exporting out of agriculture: the impact of wto accession...
TRANSCRIPT
Exporting out of Agriculture: The Impact of WTO Accession on
Structural Transformation in China∗
Bilge Erten † Jessica Leight ‡
October 4, 2017
Abstract
This paper analyzes the effect of China’s accession to the World Trade Organization in
2001 on structural transformation at the local level, exploiting cross-sectional variation in
tariff uncertainty faced by local economies pre-2001. Using a new panel of approximately
2,000 Chinese counties observed from 1996 to 2013, we find that counties more exposed
to the reduction in tariff uncertainty post-accession are characterized by increased exports
and foreign direct investment, shrinking agricultural sectors, expanding secondary sectors,
and higher total and per capita GDP. These findings are robust to a range of alternate
specifications, and to controlling for other contemporaneous reforms.
JEL Classification: F14, F16, O14, O19
∗For their comments and suggestions, we would like to thank seminar participants at Northeastern University,the Office of the Chief Economist at the State Department, Boston University, the University of California -Santa Barbara, the Stockholm Institute of Transitional Economics, and the Asian Meeting of the EconometricSociety. We would also like to thank Daron Acemoglu, David Autor, Brian Kovak, Brian McCaig, MaggieMcMillan, Mindy Marks, Daniele Paserman, Nina Pavcnik, Ivan Petkov, Dani Rodrik, and Xiaobo Zhang fordetailed comments. All errors are, of course, our own.†Department of Economics, 43 Leon Street, 312A Lake Hall, Northeastern University, Boston, MA 02115.
[email protected].‡Department of Economics, American University, 4400 Massachusetts Avenue NW, Washington, DC 20016.
1 Introduction
Over the past two decades, the explosive growth in manufacturing exports from China has
dramatically reshaped the world economy—and increasingly, heated debates about the effect of
this growth on the economies of the U.S. and Europe are reshaping politics in the developed
world as well. Between 1996 and 2013, China’s manufacturing exports increased from 3% to 17%
of total manufacturing exports worldwide, and its per capita income in purchasing power parity
terms increased from 29% to 92% of the global average.1 Moreover, a large body of empirical
work has documented that this dramatic expansion of Chinese exports had substantial and
persistent effects on manufacturing employment and wages in the U.S. and Europe.2
However, relatively little is known about the parallel effects of this expansion on local labor
markets and structural transformation in China, despite a broader literature documenting that
its manufacturing sector expanded robustly during this period, stimulated partially by export-
led growth. In this paper, we provide new evidence of the effect of positive shocks to China’s
export sector on employment, output, and value added in primary, secondary and tertiary
production at the county level; here, the primary sector includes agriculture and agricultural
extensions, the secondary sector includes manufacturing and mining, and the tertiary sector
includes services. We employ a newly assembled panel that includes a nationwide sample of
approximately 2,000 counties observed between 1996 and 2013, and utilize an identification
strategy that allows us to examine the effects of cross-sectionally varying shocks generated by
China’s accession to the World Trade Organization (WTO) in 2001.
China’s accession to the WTO was the culmination of a lengthy period of negotiations and
internal liberalization, but it did not lead to discontinuous changes in the tariff rates imposed
on Chinese exports by its trading partners, including the United States. In fact, Chinese
products had enjoyed low tariff rates in the U.S. market since 1980. However, WTO accession
did remove uncertainty pertaining to China’s Most Favored Nation (MFN) status within the
U.S. Previously, this status required annual renewal by Congress, a risky process that generated
considerable uncertainty around tariff rates. If the renewal had failed, Chinese exports would
have been subject to the much higher non-MFN rates reserved for non-market economies.3 The
U.S. permanently granted Normal Trade Relations (NTR)—a U.S. term for MFN status—to
1The years between 1996 and 2013 will be the main period of interest in this analysis. The data are drawnfrom the World Development Indicators.
2See Autor et al. (2016) for a useful review; the literature will be discussed in more detail later in theintroduction.
3For example, in 2000, the average U.S. MFN tariff was 4%, but China would have faced an average non-MFNtariff of 31% had its MFN status been revoked.
1
China in October 2000, tied to China’s accession to the WTO in December 2001 and effective
as of January 1, 2002 (Handley and Limao, 2017). This generated a sharp and sudden decrease
in the tariff uncertainty surrounding Chinese exports in a key export market. By contrast, the
status of Chinese exports in other major markets was not affected by WTO membership.
Our empirical design utilizes variation across industries in the gap between the NTR tar-
iffs permanently granted by the U.S. post–2001 and the non-NTR rates, in conjunction with
variation across counties in the composition of employment by industry reported in the 1990
census.4 The interaction of these two sources of variation generates a county-level variable,
denoted the NTR gap, capturing the exposure of local industries to tariff uncertainty prior to
WTO membership. The county is an important unit of analysis in the literature on the Chi-
nese economy, corresponding to a local labor market with defined fiscal and economic policies
(Chen and Kung, 2016; Zhang, 2006). If tariff uncertainty is a significant barrier to export-
ing, then counties characterized by larger NTR gaps should benefit more from WTO accession,
experiencing more rapid export expansion and substitution into the secondary sector post–2001.
The main results indicate that counties more exposed to tariff uncertainty prior to 2001
did experience significantly faster growth in exports, greater expansion in the secondary sector,
and more rapid increases in total and per capita GDP following WTO accession, conditional on
county and province-year fixed effects and prefecture-specific trends. Comparing two counties,
one in the 25th percentile and the other in the 75th percentile of the NTR gap, the more exposed
county shows evidence of a differential 16% increase in exports, a differential 35% increase in
secondary GDP, and a differential 15% increase in total GDP in the decade following 2001. This
export-driven expansion also has ancillary effects on other sectors: productive factors shift out
of agriculture, agricultural production and value added decline, and tertiary output expands.
Our paper is the first to exploit the post-WTO reduction in tariff uncertainty in China to
analyze sectoral reallocation and growth, and the first to analyze the impact of WTO accession
on county-level economic outcomes. It is also one of the first papers to present evidence of the
employment and GDP effects of enhanced access to advanced country markets in a developing
country context.5
4Pierce and Schott (2016b) and Handley and Limao (2017) use this empirical strategy to examine the effectsof permanent NTR status on U.S. manufacturing employment and consumer prices using industry-level data.Although our findings complement these studies, our paper differs significantly in its focus on the regional effectsof this trade policy change using detailed country-level data from China.
5Further details about relevant papers are provided below, but we should note that Facchini et al. (2016) use asimilar identification strategy in the Chinese context, but analyze only the impact on migration at the prefecturelevel, while Liu and Ma (2016) and Feng et al. (2016) analyze the effect of tariff uncertainty in China on firm-leveloutcomes. Other related papers analyzing access to advanced country markets include McCaig (2011a), McCaigand Pavcnik (2014b) and McCaig and Pavcnik (2014a) all focusing on Vietnam.
2
While a range of both internal and external reforms linked to China’s WTO membership
had a significant impact on its economic evolution in this period, we preferentially focus on
the impact of the reduced trade uncertainty in the U.S. market for several reasons. First,
the previous literature has noted that this reduced uncertainty is a major benefit of China’s
WTO accession, and had a significant impact on industrial production in the U.S.; thus, we can
reasonably expect a parallel effect in the Chinese economy. Second, the discontinuous nature
of this reform renders it conducive to analysis. Third, given that tariff uncertainty ex ante
reflects differences between NTR and non-NTR rates determined by the U.S. Congress in the
Smoot-Hawley Tariff Act of 1930, the potential for endogeneity in the NTR gap is limited.
Importantly, however, our empirical specifications all control for variation in U.S. tariff levels
during this period, as well as a range of other trade reforms implemented by both China and
the U.S., including the elimination of export licensing requirements, the reduction in barriers
to foreign investment, and the expiration of the Multi-Fiber Arrangement (MFA). In general,
variation in the level of tariffs imposed by the U.S. and other trading partners during this period
is small in magnitude relative to the potential increase in tariffs risked if China’s NTR status
had been revoked prior to WTO accession. While we show that this variation in levels has some
effect on economic outcomes, the effects of tariff uncertainty prove to be significantly larger.
In addition, there is no evidence of any significantly different trends when comparing counties
characterized by different NTR gaps prior to China’s WTO accession. When we estimate a
more complex specification evaluating the differences between high and low NTR gap counties
over time, we observe that these counties do not show any significant difference in observable
characteristics prior to 2001. The gap in their economic trajectories emerges only post–2001,
consistent with the hypothesis that the key channel is more secure access to the U.S. market.
In addition, we verify that the key results are consistent when differential trends for counties
characterized by different initial economic conditions are included. We also conduct a placebo
test utilizing export data from UNCOMTRADE that demonstrates that the cross-sectional
variation in the NTR gap is associated only with increased exports to the U.S., and does not
predict any increase in exports to other major export markets.
This paper contributes to several related literatures. First, an extensive literature analyzes
the effects of increased manufacturing exports from China on manufacturing employment and
production in developed countries; Autor et al. (2016) provide a useful overview. Autor et
al. (2013) and Acemoglu et al. (2016) exploit variation across metropolitan statistical areas in
their exposure to competition with Chinese exports. The identification strategy employed in
3
this paper is closely related to Pierce and Schott (2016b), who use industry data to analyze the
effects of diminished trade policy uncertainty on U.S. manufacturing employment. The same
authors have also presented evidence regarding the effects of Chinese import competition on
voting patterns (Pierce and Schott, 2016a) and mortality (Pierce and Schott, 2016c).6 Similarly,
Handley and Limao (2017) structurally estimate the impact of reduced trade policy uncertainty
on U.S. consumer prices. Additional research has analyzed the effect of Chinese import com-
petition on manufacturing employment in Norway, Spain, Germany and Brazil (Balsvik et al.,
2015; Costa et al., 2016; Dauth et al., 2014, 2017; Donoso et al., 2015). Our paper contributes
to this literature by documenting the extent to which the increase in Chinese manufacturing
exports had a significant impact on local labor markets and structural transformation in China.
Second, a number of studies have sought to identify the impact of trade liberalization on
the Chinese manufacturing sector, focusing on industries or firms as the unit of analysis, and
primarily analyzing variation in tariff levels. Brandt et al. (2015) demonstrate that the reduced
Chinese import tariffs following WTO accession led to significant gains in manufacturing pro-
ductivity, and Brandt and Morrow (2014) and Manova and Zhang (2012) show that reduced
tariffs have also resulted in increased access to imported inputs. Bai et al. (2017) and Khan-
delwal et al. (2013a) analyze the impact of the removal of export restrictions and MFA quotas
on export growth and manufacturing productivity at the firm level, respectively. Recent work
has also found that the diminished trade policy uncertainty following China’s WTO accession
has boosted patent applications (Liu and Ma, 2016) and stimulated entry into export-oriented
production (Feng et al., 2016).7
Although our findings complement these studies, our paper differs significantly in its focus
on structural transformation and county-level growth, as well as the channels through which
the reduction of trade policy uncertainty following WTO accession may affect these outcomes.
In particular, utilizing county-level data rather than firm-level data allows us to identify the
effects of export expansion across multiple sectors (primary, secondary and tertiary) and to
trace patterns of factor substitution across sectors. It also enables us to identify the effects of
China’s WTO accession on the extensive, growth margin, in addition to the intensive margin.
Third, our work joins a growing literature analyzing processes of structural transformation,
6In a different subliterature, a related identification strategy is employed by McCaig (2011b), though the latteranalyzes the effects of actual tariff cuts (from non-NTR to NTR levels) in Vietnam, rather than tariff uncertainty.
7A smaller literature has analyzed structural transformation in China in a broader context. Brandt et al. (2013)analyze shifts in factor market distortions in the non-agricultural economy. Marden (2015) provides evidence thatagricultural growth in the early reform period is associated with significantly faster growth in non-agriculturaloutput. Leight (2016) analyzes the effect of Bartik-style labor demand shocks on local industrialization in China.
4
in this case stimulated by export expansion. Foster and Rosenzweig (2004) estimate the impact
of shocks to the returns to agriculture in India induced by Green Revolution technology, and
find that industrial growth is fastest in areas where agricultural growth is lagging. Hornbeck and
Keskin (2015) find no evidence that positive agricultural growth generated by the construction
of an aquifer in the U.S. stimulates non-agricultural growth, while Bustos et al. (2015) present
evidence that technological innovations in the soybean sector in Brazil generate industrial growth
only when they are labor-saving. Also relevant is recent empirical work documenting that
productivity is much lower in the agricultural sectors of developing economies compared to the
non-agricultural sectors of the same economies (Gollin et al., 2014; Lagakos and Waugh, 2013).
This suggests there may be substantial gains to aggregate productivity if factors substitute from
agricultural to non-agricultural production (McMillan et al., 2014).
Finally, our study relates to the extended literature on the causal effects of trade liberaliza-
tion on a range of economic outcomes in developing countries, including poverty and consump-
tion inequality (Hasan et al., 2006, 2012; Topalova, 2010, 2007), regional labor market outcomes
(Chiquiar, 2008; Dix-Carneiro and Kovak, 2015; Kovak, 2013; McCaig et al., 2017), investment
in human capital (Edmonds et al., 2010), and fertility and child health outcomes (Anukriti
and Kumler, 2014). This literature generally analyzes domestic tariff cuts; the existing papers
evaluating the effects of expanded access to developed country export markets primarily focus
on Vietnam. Exploiting shocks generated by a bilateral trade agreement, McCaig (2011a) finds
that the U.S. tariff cuts reduced poverty in Vietnam, and McCaig and Pavcnik (2014b) and
McCaig and Pavcnik (2014a) analyze reallocation of labor between household businesses and
the formal sector. Another recent paper analyzes trade shocks linked to China’s WTO accession
on internal migration, but it utilizes only prefecture-level data (Facchini et al., 2016). Our study
contributes to this literature by presenting evidence on the employment, GDP and value added
effects of the elimination of trade policy uncertainty in a developing country context.
The remainder of the paper proceeds as follows. Section 2 provides more background on
China’s accession to the WTO and a simple conceptual framework. Section 3 describes the
data. Section 4 presents the identification strategy and the empirical results. Section 5 presents
robustness checks, and Section 6 concludes.
5
2 Background and conceptual framework
2.1 China’s WTO accession
China’s accession to the WTO in 2001 was the outcome of a lengthy and extensive negotiation
process initiated in 1986. As a member, China both received new trade access benefits and
committed to additional, liberalizing domestic reforms. However, both the benefits and the
reforms inherent in WTO accession were largely phased in gradually and did not result in
any discontinuous jumps in 2001. It is useful to highlight the most important policy changes
implemented by China as part of this process, including reduced import tariffs, the relaxation
of export licensing rules, and fewer barriers to foreign investment.
First, Chinese import tariffs had already been sharply cut prior to 2001 (from a weighted
average of over 45% in 1992 to approximately 13% in 2000). WTO accession entailed further
cuts (to approximately 7%), but these shifts were relatively small compared to the pre-accession
reforms (Bhattasali et al., 2004). Figure 1a shows the evolution of the average weighted domestic
tariff rate over time, calculated using industry-level tariffs and the share of each industry in
total Chinese imports as reported in 1996 (the first sample year). Agricultural tariffs remained
relatively high (22%) as of 2001 and required further cuts to 17.5% by 2004, with deeper
cuts for agricultural products prioritized by the U.S. (e.g., corn). In addition, sanitary and
other non-tariff barriers to U.S. exports of citrus, meat and grains were eliminated when China
accepted U.S. inspection standards, and American companies were also allowed to freely trade
agricultural products within China (Cheong and Yee, 2003).
Second, restrictions on direct exporting were substantial prior to WTO accession, though
variable by industry, while firms that were not granted licenses to export directly were required
to export via partners. In 2000, slightly more than half of the large firms observed in annual
surveys of large industrial enterprises were not permitted to export directly, but all firms were
allowed to export freely by 2004 (Bai et al., 2017). Third, prior to WTO accession, China had
generally implemented relatively attractive policies to draw in foreign investment. However, for-
eign firms were subject to performance requirements, including criteria related to local content,
technology transfers, and investments in research and development. These requirements were
eliminated following China’s accession to the WTO, facilitating a more rapid inflow of foreign
investment (Long, 2005).
What about changes in the tariffs imposed by trading partners? Figure 1b shows fluctuations
in tariffs over time for China’s most important trading partners: the NTR tariffs imposed by the
6
U.S. and the average tariff rates imposed on Chinese exports by the European Union, Japan,
Korea, and Taiwan. On average during this period, the U.S. is the destination for approximately
20% of Chinese exports, followed by the European Union at 17%, Japan at 12%, Korea at 5%
and Taiwan at 2%. We again construct these rates as weighted averages of industry-level tariffs,
utilizing the shares of total exports constituted by each industry’s output in 1996 as weights.
The estimated tariffs imposed by Korea are highest (averaging between 8% and 10%), but show
no significant trend. Tariffs imposed by the U.S. and Taiwan decline gradually, and the tariffs
imposed by Japan and the EU are roughly constant. In all cases, there is no evidence of any
dramatic shifts in tariff rates at the point of China’s WTO accession.8 Despite their gradual
nature, however, all of the preceding shifts in trade policy are relevant in understanding the
evolution of local economies during this period, and we will control for these shifts in our
empirical specifications.
Importantly, there was a discontinuous jump in one important dimension of China’s market
access in 2001: the tariff uncertainty faced in the U.S. market. Prior to WTO accession,
the United States granted China NTR tariff rates on a discretionary basis subject to annual
congressional renewal. Failure of that renewal would have triggered the imposition of much
higher tariffs, originally set by the Smoot-Hawley Act, and designated for non-market economies.
Hence, although the tariff applied to Chinese imports remained low because China’s NTR status
was never withdrawn, the required annual approval generated considerable uncertainty. Using
media and government reports, Pierce and Schott (2016b) document that firms perceived the
annual renewal of MFN status as far from guaranteed, particularly in periods of political tension
in the early 1990s.9 The CEOs of 340 firms stated in a letter to President Clinton that the
“persistent threat of MFN withdrawal does little more than create an unstable and excessively
risky environment for U.S. companies considering trade and investment in China, and leaves
8In Figure A1 in the Appendix, we provide an alternate representation of the evolution of both domestic tariffsand trading partner tariffs over time, utilizing county-level employment weights provided by the 1990 census tocalculate a county-level weighted average tariff and then reporting the mean weighted county-level tariff by yearover time. (These average county-level tariffs will subsequently be employed as control variables in the regressionsof interest.) The graphs are largely similar, except that the tariffs imposed by Korea on Chinese imports appearmuch higher, reflecting Korea’s extremely high tariffs on agricultural exports from China.
9Anecdotal evidence from the Chinese media has emphasized that China’s WTO accession “will help buildconfidence among investors at home and abroad, especially among United States investors, because China cur-rently faces the issue of maintaining its Most Favored Nation trading status every year” (Shanghai SecuritiesNews, 1999). Chinese companies have also expressed that “[they] can enjoy multilateral Permanent Most Fa-vored Nation status among the Member States of the WTO, so as to actively explore and enter the internationalmarket and participate in international economic competition” (Jiangxi Paper Industry Co. Ltd., 2000). Chinesenewsletters described the U.S.’s decision to sever the ties between China’s MFN status and human rights recordas having “removed a major issue of uncertainty”; in addition, the renewal of China’s MFN status would encour-age investment and re-exports by “removing the threat of potential losses that would have arisen as a result ofrevocation” (South China Morning Post, 1994).
7
China’s booming economy to our competitors” (Rowley, 1993).
In October 2000, Congress passed a bill that granted permanent NTR status to China,
effective as of January 1, 2002. The EU had granted China permanent NTR status much
earlier (effective in 1980); thus, China did not face any tariff uncertainty in this market either
before or after its WTO accession (Pierce and Schott, 2016b). The permanency of China’s NTR
status in other markets is ambiguous, but the descriptive evidence generally suggests there were
no dramatic changes in the status of China’s exports to other markets during this period, and
analysts have noted that WTO members other than the U.S. had already provided China with
permanent MFN status prior to its accession to the WTO (Rumbaugh and Blancher, 2004).
In Figure 2, we provide some preliminary graphical evidence consistent with the hypothesis
that WTO accession represented a major turning point in the Chinese economy. We can observe
that employment in the primary (agricultural) sector declined rapidly between 1990 and 2015,
while secondary and tertiary employment increased (Panel A). At the same time, the compo-
sition of output also shifted dramatically toward the secondary and tertiary sectors (Panel B).
Both panels show some evidence of a trend break around 2001, marking China’s accession to
the WTO. Figure 3 similarly highlights the dramatic expansion of China’s exports to the U.S.
and the concomitant rise in its import penetration ratio in the U.S. market post–2001.10
Again, a number of policy shifts during this period shaped economic outcomes. However,
we will preferentially focus on reduced trade uncertainty given that the previous literature has
highlighted this shift had a major impact on the U.S. market, and given the discontinuous
nature of the reduction in uncertainty. We will also present evidence that while the other
reforms implemented during this period had a meaningful impact on local economic outcomes
in China, the effect of reduced tariff uncertainty generally proves to be largest in magnitude.
Our analysis allows us to separately identify the impact of tariff uncertainty vis-a-vis levels by
exploiting the fact that tariff uncertainty varies only comparing the pre and post period, and is
proxied by the difference between low tariff rates and the counterfactual high rates specified by
the U.S. tariff schedule. By contrast, realized tariff levels imposed by both the U.S. and other
trading partners vary continuously over time. Further details are provided in section 3.2.
10China’s import penetration in the U.S. market is defined as U.S. imports from China divided by the total U.S.expenditure on goods, measured as the gross U.S. output plus U.S. imports minus U.S. exports. The last threeseries are drawn from the World Development Indicators. While Chinese imports from the U.S. also increased inthis period, the rate of increase is much smaller compared with the growth in China’s exports.
8
2.2 Conceptual framework
The reduction of tariff uncertainty can affect county-level economic outcomes through several
channels. First, a reduction in tariff uncertainty creates incentives for Chinese firms to increase
their exports to the U.S. market. A large literature has established that price uncertainty (in
this case generated by tariff uncertainty in the destination market) generates an option value
of waiting, decreasing investment (Bernanke, 1983; Dixit, 1989; Bloom et al., 2007). When
tariff uncertainty is reduced, firms facing positive demand in the destination market, primarily
manufacturing firms, have a greater incentive to make irreversible investments required to enter
foreign markets (Handley and Limao, 2015; Handley and Limao, 2017). Given that industries
differ in their exposure to tariff uncertainty, firms in industries with greater exposure ex ante
will face a greater decline in the option value of waiting post-WTO accession. Exports from
these industries, and counties with a greater concentration in these exposed industries, will
differentially increase.
Second, a reduction in tariff uncertainty induces U.S. firms to increase foreign direct invest-
ment (FDI) into China, as again the option value of delaying investment declines. In addition,
export-oriented industries in China are generally characterized by high FDI, as foreign investors
producing for export have benefited from a variety of preferential policies, including the exemp-
tion of imported components from import duties (Zhang and Song, 2000) and the establishment
of preferential zones that offer reduced taxes on profits and other benefits (Cheng and Kwan,
2000). Accordingly, a growing export sector can be expected to attract increased FDI, and these
effects would be particularly large in industries and counties more exposed to tariff uncertainty
ex ante and those industries facing non-trivial foreign demand, primarily in manufacturing.
Third, another channel through which a reduction in tariff uncertainty may affect county-
level economic outcomes is the reallocation of productive factors across sectors. Increased
demand for exports and increased FDI in the secondary sector will increase the returns to
capital and labor, leading factors to flow into this sector (Acemoglu et al., 2016). We denote
this the local reallocation effect. On the other hand, an increase in exports and FDI at the
county level generates a positive local income effect and an increase in local demand, benefiting
producers of non-tradables, as well as any producers of tradables that sell partly to the local
market. We describe this as the local demand effect. If there is some input in the tradable sector
that is not mobile across sectors, the local demand effect will dominate the reallocation effect
for the non-tradable (tertiary) sector, leading to its expansion (Kovak, 2013).11 In addition, if
11The existing literature analyzing the response of U.S. local labor markets to Chinese trade shocks also find
9
preferences are non-homothetic, a positive local income effect will shift consumption away from
agricultural and agricultural-derived goods, reinforcing the reallocation of productive factors
toward the secondary sector (Gollin et al., 2014).
By examining economic outcomes at the level of counties, or local labor markets, we are
able to capture both the direct effect of reduced uncertainty on the expansion of sectors that
benefit from increased exports and increased FDI, as well as the indirect effects generated by the
reallocation of productive factors and the expansion of local demand. Moreover, the reduction
in tariff uncertainty may have disproportionate effects on counties with certain baseline char-
acteristics. Since capital investments are generally irreversible, counties with an initially higher
concentration of capital-intensive industries are likely to respond more robustly to the reduction
in tariff uncertainty. Similarly, the effect of reduced tariff uncertainty is likely to be larger for
counties that specialize in industries exporting a higher proportion of their output to the U.S.
ex ante. Finally, counties that include special export zones, and thus host a particularly high
concentration of export-oriented firms, may show a particularly large response to the reduction
in tariff uncertainty. We will also test these hypotheses in the empirical analysis.
3 Data
The empirical analysis incorporates three sources of data: county-level economic outcomes, the
county-level NTR gap, and other policy shifts. We will discuss each data source in turn.
3.1 County-level data
The main outcomes of interest are economic indicators at the county level reported by provin-
cial economic yearbooks. Each year, every province in China publishes a statistical yearbook,
primarily reporting economic indicators for the full province or for larger aggregate units such
as prefectures. However, most provincial yearbooks also include some economic indicators re-
ported at the county level. These data were compiled and digitized for every year available
between 1996 and 2014. (Each yearbook reports data from the previous year; thus, 2013 is
the final year observed in the data.) To the best of our knowledge, this study is the first to
construct a comprehensive county-level panel of economic outcomes for this time period.
Only one limitation is imposed on the sample. We exclude provincial-level autonomous
regions: Tibet, Xinjiang, Ningxia, Inner Mongolia, and Guangxi, as well as the island of Hainan,
that the local demand effect dominates local reallocation effects (Autor et al., 2013; Acemoglu et al., 2016).
10
for which data is generally unavailable. Otherwise, all counties that can be matched between the
1990 county census and the provincial yearbooks are included. Aggregated to the county level,
the 1990 census reports data on 2104 units that are (approximately) at the county level in the
provinces of interest; of these units, 86%, or 1805 counties, can be matched to the yearbooks.12
The county-level panel includes information on exports; GDP, value added and employment
by sector; and detailed information about investment in agriculture. GDP, value added and
employment are reported for the primary, secondary, and tertiary sectors. Again, the primary
sector includes agriculture, fishing, and forestry; the secondary sector includes manufacturing
and mining; and the tertiary sector includes services. Exports, GDP and value added are re-
ported in millions of yuan, and per capita GDP is reported in yuan. The nominal figures for
GDP and value added reported in the provincial yearbooks are deflated using World Bank defla-
tors. Additional variables capturing investment in agriculture include cultivated area (reported
in thousands of hectares), agricultural machinery used (reported in 10,000 kilowatts), and grain
and partial cash crop output (reported in thousands of tons).13
Summary statistics are reported in Table 1. The average population in the sampled counties
and years is approximately 500,000. More than half of reported employment is in the primary
sector, followed by the tertiary and secondary sectors. Per capita income is approximately 10,000
yuan or $1300. The largest share of GDP is constituted by the secondary sector, followed by
the tertiary and primary sectors.
Missing data Data is missing from the county-level panel for two reasons: counties cannot
be matched between the census and the provincial yearbooks, and counties are matched to the
yearbooks but specific indicators are not available. Here, we will briefly discuss each case; a
detailed discussion can be found in Section A1.1 in the Appendix.
First, some counties that are observed in the census do not appear in provincial yearbooks.
These are disproportionately counties that are part of larger, prefecture-level cities, as some
provinces omit data for these areas. Accordingly, any bias due to missing counties will orient
the sample toward rural areas that are not already fully industrialized. The differences between
12The 1990 census has one unusual characteristic that differentiates it from subsequent census rounds (2000and 2010) and from the provincial yearbooks: data for prefecture-level cities are reported only at the prefecturelevel, not for the constituent county-level units. In some cases, provincial yearbooks report data for these county-level units of prefecture cities. Accordingly, a single census observation can in these cases be linked to multiplecounty-level observations in subsequent waves of yearbook data.
13The production of cash crops is calculated as the sum of the production of meat and edible oils, the mostcommonly reported cash crops. This is clearly an incomplete measure of cash crop production, but allows us togenerate some evidence about evolution of non-staple cultivation.
11
counties observed and not observed in provincial yearbook data are summarized in Table A1
in the Appendix, in which we estimate a series of specifications regressing county covariates as
observed in the 1990 census on a dummy for missing, conditional on province fixed effects. The
results suggest that counties missing from the sample are characterized by larger populations,
higher levels of education, and a greater concentration of labor outside of agriculture.
Second, for those counties that are observed in provincial yearbooks, different provinces in
different years opt to report different indicators at the county level in their yearbooks. As a
result, the number of observations varies significantly for different variables, as evident from the
summary statistics. The indicators that are reported most infrequently include employment at
the sector level and exports, while indicators reported near-universally include gross domestic
product, total employment, population, primary and secondary value added, and measures of
agricultural inputs and production.14 (For each variable presented in Table 1, we also note the
number of counties reporting any data for that variable. This figure ranges between 1400 and
1700, excepting exports and employment at the sectoral level.)
We also present further evidence in Table A2 in the Appendix that the number of obser-
vations for the key variables of interest is in general lower for more populous counties, and
higher for those that are more agricultural and have a lower proportion of employment outside
the primary sector. This is again consistent with more urban and industrialized counties being
generally underrepresented in the sample. We will subsequently demonstrate that the primary
results are all robust to controlling for patterns of selection into the sample. In addition, we will
present evidence around the evolution of exports and secondary employment — key outcomes
of interest that are infrequently reported in the county-level data — drawing on an additional
survey of large-scale manufacturing firms.
3.2 County-level NTR gap measure
Our empirical analysis seeks to identify the effect of the substantial reduction in tariff un-
certainty in the U.S. market that China experienced following its accession to the WTO. To
estimate the impact of China’s permanent NTR status, we define the NTR gap at the industry
level for each of the 39 subsectors of tradable production represented in the census data.
NTRGapi = NonNTRRatei −NTRRatei (1)
14In particular, a strong positive correlation exists between the probability of reporting any data on exportsales value and county-level GDP, and five relatively poor provinces (Shanxi, Sichuan, Guizhou, Shaanxi, Gansu,and Qinghai) report almost no data on exports
12
The NonNTRRatei is the higher tariff rate that would have applied if the U.S. Congress had
revoked China’s annual NTR status for industry i, and the NTRRatei is the lower tariff rate
guaranteed by permanent NTR status.
The industry-level NTR gap data were constructed by Pierce and Schott (2016b) using ad
valorem equivalent NTR and non-NTR rates. The NTR gap for industry i is the average NTR
gap across the four-digit ISIC Revision 3 tariff lines belonging to that industry. Throughout
the empirical analysis, we use the NTR gaps for 1999, two years before the U.S. granted China
permanent NTR status.15 We manually match the industry categories in ISIC Revision 3 to the
industry categories reported in the Chinese employment data, and Table A1 in the Appendix
provides the details associated with this matching.
We then construct a county-level NTR gap measure equal to the weighted average of industry
gaps, where the baseline composition of employment by industry prior to WTO accession is used
to construct the weights. More specifically, we utilize the employment data reported in the 1990
census to calculate the share of tradable employment by industry in each county, interacting
the NTR gap faced by industry i with each industry’s county-specific employment share.
NTRGapc =∑i
empshare1990ic ×NTRGapi (2)
Given that each county’s sectoral composition prior to WTO accession is used to construct
the employment shares, the NTR gap does not reflect endogenous changes in employment
composition that are driven by reduced trade policy uncertainty. Counties characterized by a
larger NTR gap experience a greater reduction in trade policy uncertainty post–2001, and thus
ceteris paribus should show greater expansion in export-oriented industries. Permanent NTR
rates were effective for China as of January 1, 2002, and thus our analysis characterizes all years
from 2002 onward as the post-reform period.
In addition, we preferentially employ the employment shares observed in the 1990 census
rather than the 2000 census to minimize potential endogeneity in employment composition. We
hypothesize that by 2000, counties with more informed leaders or enterprises with more foresight
may have already shifted toward subsectors that were less exposed to trade policy uncertainty.
This would generate some correlation between county-level unobserved characteristics and the
size of the county NTR gap. We will subsequently demonstrate that the results are robust to
the use of 2000 employment weights, and are also consistent when the employment shares are
15We follow Pierce and Schott (2016b) in utilizing the 1999 NTR gaps. These NTR gaps are almost identicalto those in 2000 or 2001; accordingly, the results are robust to the use of data from other years.
13
recalculated with respect to total employment (including non-tradable employment).16
Table A5 in the Appendix summarizes the NTR gap observed for each industry. The highest
NTR gaps are observed for textiles, garments, other manufacturing, medical and pharmaceutical
products, and furniture manufacturing; the lowest NTR gaps are observed for mining products
and agricultural output. At the county level, the average NTR gap is .123 with a standard
deviation of .043. Approximately 5% of counties face NTR gaps of more than 20%.
Figure 4 shows a histogram of the NTR gap at the county level. While there is some
evidence of outliers, we will demonstrate that the primary results estimated in Section 4.1 are
robust to winsorizing the NTR gap. Figure A2 in the Appendix shows a map of cross-country
variation in the NTR gap, utilizing the residuals after the NTR gap is regressed on province
fixed effects. Overall, there is substantial variation in exposure to reduction in tariff uncertainty
across Chinese counties.
3.3 Other policy changes
In the main empirical analysis, we also consider a number of other policy changes in China
and the U.S. to isolate the impact of China’s accession to the WTO. In particular, we exam-
ine whether other policy shocks could be the cause of the structural change that China has
experienced over the past decade. Other policy shocks may constitute plausible alternative
explanations if their timing coincides with China’s WTO accession and if these shocks would
disproportionately affect counties that are more exposed to reduced tariff uncertainty post–
2001. As previously noted, major domestic reforms in this period included lower import tariffs,
the elimination of import licensing requirements, and reduced restrictions on FDI.
In our regressions, we use data on China’s import tariffs from the WITS–TRAINS database,
data on export licensing requirements from Bai et al. (2017), and data on the nature of con-
tracting from Nunn (2007) to control for these policy changes. The data on the nature of
contracting provide a measure of the proportion of intermediate inputs employed by a firm
that require relationship-specific investments by the supplier; counties with high concentrations
of industries characterized by different contracting methods may be differentially affected by
reductions in barriers to foreign investment. For each of these variables, we construct a county-
16Data on GDP, revenue, and export value per subsector are not available in any year; accordingly, weightscan only be constructed using employment data. Constructing measures of exposure to trade shocks usingemployment weights is common in the literature, and has been theoretically derived by Kovak (2013) as thecorrect measure of trade exposure. Employment weights are also employed by Topalova (2007, 2010), McCaig(2011a), Kovak (2013), and Autor et al. (2013) in analyzing the effects of trade exposure on poverty and locallabor market outcomes in regional labor markets in India, Vietnam, Brazil, and the United States, respectively.
14
level weighted average from the industry-level source data using employment weights from the
1990 census.17
We also control for policy changes in the U.S., including the time-varying NTR rate itself, for
which we construct an industry-weighted county average. An additional important policy shift
during this period was the elimination of textile and clothing import quotas in 2002 and 2005
as part of the global MFA. We employ data on MFA quotas from Khandelwal et al. (2013b),
and follow their methodology to construct a measure of the degree to which industries’ quotas
were binding under the MFA by calculating the import-weighted average fill rate. The fill
rates represent the ratio of actual imports to allowable imports under the quota; thus, a higher
value indicates greater exposure to MFA quota reductions. Using these industry-level data, we
construct a county-level MFA variable, where greater values represent greater exposure to quota
reductions and thus greater benefits from the policy shift.
4 Empirical results
In this section, we first analyze the baseline specification focusing on pre-post differences. Next,
we present evidence that counties with high and low NTR gaps are characterized by parallel
trends prior to 2001, but diverging economic trajectories post-accession. We also demonstrate
that the results are robust to a number of additional specifications, and draw on additional data
from a survey of large-scale firms.
4.1 Baseline specification
First, we use a difference-in-difference specification to analyze the effect of reduced trade policy
uncertainty on county-level economic outcomes. More specifically, we examine whether the
trajectory of economic outcomes in counties characterized by relatively large gaps between NTR
tariff rates and non-NTR rates is different following China’s accession to the WTO in 2001. The
sample includes annual county-level data from 1996 to 2013; all the dependent variables have
the top and bottom 2% of observations trimmed from each year to eliminate outliers.18
17Since the industry categories for the export licensing and contract intensity variables are available for SICcategories, these categories are manually matched to the census employment categories. The industry classifica-tion for the import tariff data is available in ISIC Revision 3, the same source utilized to construct the NTR gapvariable. Table A1 in the Appendix provides the details associated with the matching.
18This trimming process is implemented separately for urban and non-urban county units. The results are alsoconsistent if only the top and bottom 1% are trimmed or if no observations are trimmed; the results for thesespecifications are reported in Table A9 in the Appendix. We also report in Panel C of this table results using theoriginal sample but winsorizing the NTR gap at the 1st and 99th percentile; these results are likewise consistent.
15
We employ ordinary least squares (OLS) to estimate the following specification:
Ycfpt = β1Postt ×NTRGapcfp + X ′cfptθ + γpt + νf t (3)
+ Urbcfp × (γpt + νf t) + δc + εcfpt
The dependent variable is observed in county c in prefecture f in province p in year t. The
independent variable is the interaction of the county-level NTR gap, standardized to have a
mean of zero and a standard deviation of one, with a post–WTO dummy, equal to one for 2002
and subsequent years.19
The specification also includes a number of additional controls denoted X ′cfpt. This includes
the interaction of the post dummy and a time-invariant dummy capturing whether the county
is characterized by industries with high contract intensity.20 We also control for time-varying
shocks: the industry-weighted MFA quota fill rate for county-produced goods, the industry-
weighted domestic import tariff rate, the industry-weighted percentage of local firms licensed
to export, and the industry-weighted NTR tariff rates. (All variables capturing other changes
in trade policy during this period are also included in the specifications estimated in Pierce and
Schott (2016b); we will demonstrate that the results are consistent when estimated without these
additional controls.21) The specification also includes province-year fixed effects, prefecture-
specific trends, and county fixed effects. The time-varying fixed effects are interacted with an
urban dummy to allow for differential trends in urban areas, and standard errors are clustered
at the county level.22
The results of estimating equation (3) are reported in Table 2; for concision, only the
coefficient β1 is reported. (The full set of coefficients is reported in Tables A6 through A8 in
the Appendix, and will be discussed subsequently.) To analyze the magnitude of the effects, we
will consider the effect of a county moving from the 25th to the 75th percentile of exposure to
tariff uncertainty, an increase of .48 standard deviations in the standardized NTR gap. In Panel
19The county-level NTR gap is omitted, given the inclusion of county fixed effects; similarly, the post dummyis collinear with province-year fixed effects.
20Specifically, this dummy is equal to one if the weighted average of industry contract intensity is above themean.
21There are some differences between our specification and that employed in Pierce and Schott (2016b). Theyinclude the contract intensity variable in linear form and use the import tariff and export licensing variablesto construct differences over time that interact with the post–WTO dummy. They also include other controlvariables for baseline capital and skill intensity and the use of high-technology products that are unavailable inour data. We also interact the domestic tariff rate with the post dummy, given that most of the tariff reductionsobserved in China were implemented prior to its WTO accession.
22This dummy variable is equal to one if the county name includes the “shi” (i.e., city) suffix in 1990. Approx-imately 19% of the counties are designated as urban.
16
A, we observe that this increase would lead to an increase in exports of approximately 16% in
the post–2001 period.23 There is also evidence of a relative increase in secondary GDP of 35%,
a relative increase in tertiary (non-tradable) output of 19%, a relative increase in total GDP of
15%, and a relative increase in per capita GDP of 5%. No significant effects are observed for
primary output.24
Panel B reports the employment results; again, employment data are available for a relatively
small sample, and the results are thus more noisily estimated. When comparing counties in the
25th and 75th percentiles of the NTR gap, we observe an increase in secondary employment
of approximately 6%, and weak evidence of a decline in primary employment. Agricultural
employment, a closely related measure reported for a larger sample, shows a significant decline
of a magnitude symmetric to the increase in secondary employment. There is no shift in
tertiary or total employment, but we observe a relative increase in population of approximately
3% in counties that were ex ante more exposed to tariff uncertainty, suggestive of in-migration
to counties where export-driven manufacturing is growing. The population increase is also
consistent with the previous evidence that the increase in GDP per capita is small relative to
the increase in total GDP.
Finally, Panel C reports the results for agricultural investment and value added. We observe
declines in the utilization of agricultural machinery, grain production, and value added in the
primary sector of between 1% and 4%. (There is no effect observed for cash production, though
this evidence should be interpreted cautiously given that cash crops include only meat and
edible oils.) Secondary value added increases by 11%.25 Given the small sample of sectoral
employment, it is not possible to estimate results for value added per worker as a proxy for
productivity; however, we will subsequently present results for value added per worker in the
secondary sector using data from the large-scale firm survey.
Taken together, these results suggest a clear pattern. Counties with high concentrations of
industries exposed to large gaps between NTR and non-NTR tariffs show evidence of signif-
icantly more expansion in the secondary sector following China’s WTO accession—a pattern
evident in increased employment, higher GDP, and higher value added—and this growth gen-
erates an increase in local GDP and GDP per capita. There is also some evidence of greater
23The magnitudes reported in this discussion are all calculated multiplying the observed coefficients by .48 anddividing by the reported mean at the bottom of the relevant panel.
24Unfortunately, the county-level data do not report any information on imports. Data on imports are providedat the provincial level; analyzing the effect of the post-NTR gap interaction in a parallel specification estimatedwith data at the province-year level reveals only weak evidence of an increase in imports.
25All of these estimates are consistent if province-year fixed effects are replaced with year fixed effects.
17
contraction in the agricultural sector as productive factors substitute into secondary production.
The coefficients for the full set of control variables are reported in Tables A6 through A8
in the Appendix. First, the coefficient for the interaction between the post dummy and the
high contract intensity dummies is generally negative for measures of secondary investment
and output. This finding suggests that industries characterized by more relationship-specific
contracting for inputs benefit less from WTO accession, presumably because these industries
had experienced more foreign investment ex ante (a comparative static that can be substantiated
in this data). Second, there is more rapid growth in counties that benefit more from MFA quota
reductions, and slower growth in counties more exposed to a decline in domestic tariff rates and
an increase in competition from imports. However, the effects of import tariff reduction are
observed primarily prior to 2001, consistent with the evidence previously presented around the
timing of the tariff cuts. The coefficients on export licensing and the time-varying NTR rate
vary in significance. These patterns are consistent with the hypothesis that, while other trade
reforms in this period were relevant for the evolution of county-level outcomes, no other policy
shift had a positive effect on county-level expansion of exports and secondary production as
large as that produced by reduced uncertainty in the U.S. market.26
In Table 3, we re-estimate the NTR gap using a number of alternate strategies to evaluate
the robustness of these results; for brevity, we focus only on exports and GDP. (In each case, the
control variables calculated as county-level weighted averages are also re-estimated.) In Panel
A, we construct the NTR gap utilizing the employment data reported in the 2000 census to
construct employment weights rather than utilizing the 1990 weights. The results are generally
comparable, although the estimated coefficients for exports, GDP and per capita GDP are
larger. The use of 2000 employment weights introduces two potential sources of bias: areas
already industrialized by 2000 will generally have larger NTR gaps, while industrialized areas
that are more strategic in investing in industries characterized by less tariff uncertainty may
have lower NTR gaps. The former phenomenon will lead to upward bias in the estimates of the
NTR gap if already-industrialized counties continue to expand more rapidly, and this upward
26Two seeming anomalies can be observed in the signs of the control variables. First, the proportion of firmslicensed to export is negatively correlated with GDP; second, the NTR tariff rate is positively correlated withGDP. In the cross-section, we observe the expected positive correlation between the proportion of firms exportingand county-level GDP prior to 2004 (when export licensing requirements were eliminated). However, when countyfixed effects are included, counties that show larger increases over time in export licensing are, mechanically, thosewith initially lower levels of export licensing, given that the maximum value for this variable is one. These countieswith low initial export license levels are also characterized by slower GDP growth. In addition, counties with alarger NTR gap are, on average, characterized by a higher NTR tariff rate (and an even higher non-NTR rate),producing the observed positive correlation between the NTR tariff rate and economic outcomes.
18
bias does seem to be evident in these specifications.27
In Panel C, we construct the NTR gap by weighting each subsector with respect to total
employment, assigning a zero weight to the tertiary (non-tradable) sectors. In our main spec-
ification, we estimate the NTR gap without considering the relative size of the services sector,
weighting employment with respect to total employment in tradable sectors; this methodology
is recommended by Kovak (2013), though earlier papers in the trade liberalization literature
assign the non-tradable sector a weight of zero.28 Using this alternate strategy to re-calculate
the NTR gaps and re-estimate equation (3) yields consistent results.
As previously noted, in general the gap between NTR tariffs and non-NTR tariffs is relatively
low for agricultural products compared with that for industrial products; this raises the potential
challenge that the observed growth in high NTR gap counties post–2001 may primarily reflect
more rapid growth for already more heavily industrialized counties. Another related source of
bias may stem from the fact some of the highest NTR gaps are observed for textiles and garment
manufacturing, sectors that also benefited considerably from the relaxation of the MFA quotas.
While the main specification includes controls for county-level variation in quotas, bias could
be introduced by any shocks to textile production that are not captured by this variable.
We will address both points by implementing a similar strategy: including additional control
variables for employment shares in different sectors interacted with year fixed effects. First, we
calculate the share of employment in the secondary and tertiary sector as observed in the 1990
census, construct separate quartile dummy variables for each employment share, and include
interactions between the quartile dummy variables and year fixed effects in the primary speci-
fication. Second, we use the employment shares in the five sectors characterized by the largest
NTR gaps (textiles, garments, other manufacturing, medical and pharmaceutical products, and
furniture manufacturing), again construct five sets of quartile dummy variables, and interact
these variables with year fixed effects. (We use quartile dummy variables rather than the con-
tinuous variables to flexibly allow for non-linear effects of variation in employment shares; the
results are also consistent if we simply employ the linear variable.) The results are reported in
Panels C and D of Table 3, and are consistent and in fact somewhat larger in magnitude.
An alternate strategy would entail re-estimating the NTR gap while omitting certain sectors.
For example, we can estimate a secondary-only NTR gap, excluding all agricultural subsectors
27The number of observations increases slightly, as some county codes can be matched to the 2000 census butcannot be matched to the 1990 census. These results are also consistent if we employ the mean of sector weightsas observed in the 1990 and 2000 censuses.
28This strategy has been widely used; see, for example, Autor et al. (2013), McCaig (2011a), Topalova (2007),and Topalova (2010).
19
from the calculation of the NTR gap and calculating only the weighted average for secondary
subsectors. In addition, we can estimate a “no high gap” NTR variable, excluding the five
sectors characterized by the largest NTR gaps. The primary coefficients are also consistent
if the main specification is re-estimated employing these alternate variables; these results are
reported in Table A10 in the Appendix.
There is also substantial expansion in China’s agricultural imports from the U.S. during this
period, particularly in cotton and soybeans.29 We can utilize data from the 2000 World Census
of Agriculture (FAO/IIASA) to analyze the cross-sectional correlation between the NTR gap
and the proportion of area sown in soybeans and cotton. In general, this correlation is negative,
suggesting that areas experiencing more export-driven growth are less subject to competition
from imports. If we re-estimate the main specification including an interaction term between
high cotton and soybean production (a dummy for the fraction of sown area devoted to cotton
and soybeans being above the median) and the NTR gap, the interaction terms are small in
magnitude and generally insignificant, as reported in Panel F of Table 3.30 Accordingly, com-
petition from imports is not a channel that seems to be of first-order importance in generating
the observed substitution away from agriculture.
Variation in sample size Again, the number of observations fluctuates in the main specifi-
cations because many provincial yearbooks do not report specific indicators of interest (partic-
ularly employment at the sectoral level and exports). We report in each panel the number of
unique counties observed in the sample for each variable; the number of counties ranges between
1000 and 1700, with the exception of the employment results, for which the sample includes
fewer than 400 counties. We will subsequently estimate results derived from a survey of large
firms that will enable us to corroborate the observed patterns for secondary employment and
exports for the full set of counties observed in the primary analysis.
In addition, we present in Section A1.1 in the Appendix a number of additional specifications
exploring whether the results are robust to selection into the sample, including imposing a
sample restriction to only county-years that report export data. We observe consistent results
across a number of different specifications controlling for selection into the sample, suggesting
that missing data is not a significant source of bias.
29Figure A3 in the Appendix shows the evolution of China’s agricultural imports over time.30The specification also includes interactions between dummies for each quartile of the cotton and soybean
fraction variable, measured at the prefecture level, and year fixed effects.
20
4.2 Evidence about timing
Given that we attribute the observed patterns to the reduction in tariff uncertainty following
China’s accession to the WTO in 2001, a more demanding test of the assumptions of the
difference-in-difference specification can be conducted by evaluating the correlation between
the variables of interest and the NTR gap prior to 2001. To implement this test, we estimate a
more complex specification, in which we interact the NTR gap with a series of dummy variables
for two-year intervals. (A single dummy variable captures the three-year pre-treatment interval
for the period 1997–1999.) Dummy variables for the years prior to 1997 are omitted, rendering
1996 and the small sample of pre-1996 observations the reference period. The specification of
interest can thus be written as follows, including the same control variables reported in the
simpler specification.
Ycfpt =2013∑
y=1997
βy11{y = t, t+ 1} ×NTRcfp (4)
+ X ′cfptθ + γpt + νf t+ Urbcfp × (γpt + νf t) + δc + εcfpt
The results of estimating equation (4) are reported in Table 4, employing four of the main
variables (exports, secondary output, total and per capita GDP). We observe that the coeffi-
cients for the NTR gap prior to 2002 are uniformly insignificant and generally small in mag-
nitude. (Given the limited data reported for exports prior to 1997, the dummy variable for
1997–1998 is omitted in this specification.) In particular, the absence of any significant effect in
2000–2001 is consistent with the evidence presented in Handley and Limao (2017) that China’s
new tariff status was not implemented until 2002.
However, following China’s WTO accession, the magnitudes of the coefficients for the NTR
gap variable are generally increasing over the subsequent decade; the coefficients are positive
and significant from 2004 onward, with the exception only of the coefficient for exports for
2012-13. This evidence is consistent with the hypothesis that the NTR gap is uncorrelated with
any variation in county outcomes prior to China’s WTO accession, but highly predictive of the
economic trajectories observed in the same counties post–2001.31
The coefficients are presented graphically in Figure 5. Again, we can observe that no sig-
nificant relationship exists between the NTR gap and the outcomes of interest prior to 2001
31No data for secondary and tertiary GDP are available after 2011; thus, the coefficients for the dummy variablefor 2012–2013 are missing in the specifications using these variables.
21
and that the difference-in-difference coefficients increase steadily after 2001. The pattern of an
effect that is consistently positive after 2001, but growing slowly in magnitude, is also consis-
tent with the parallel evidence presented by Pierce and Schott (2016b) in Figure 4, showing a
gradual decline in manufacturing employment in the U.S. over the same period. We can also
test whether the estimated coefficients β1 are equal when compared across the pre-treatment
period (the dummy variables for 1997–1998 and 1999–2001) and the post–2001 period. While
the coefficients for 2002–2003 are noisily estimated, the estimated coefficients for the post–
2003 period are significantly different from the estimated coefficients for the two pre-treatment
dummy variables in all but two cases.32 The remaining 31 pairwise tests yield p-values that are
significant at the 5% level, and all but two are significant at the 1% level.
To sum up, these results are consistent with the hypothesis that the observed divergence in
economic trajectories of counties subject to different gaps between NTR and non-NTR tariffs
following China’s WTO accession is primarily due to increased access to the U.S. market, leading
to an increase in exports. These patterns first emerge in the early part of the post–2001 period,
but they become steadily more pronounced over the subsequent decade.
4.3 Alternate specifications
We report a number of alternate specifications evaluating the robustness of these results in
Table A11 in the Appendix. In Panel A, we estimate the baseline specification including only
province-year and county fixed effects and prefecture-specific trends. In Panel B, we include
the full set of controls and weight each county observation by its 1990 population.33 In Panel
C, a full set of interactions between year fixed effects and a dummy variable for each quartile
of initial GDP are added.
In Panel D, we characterize counties based on the proportion of the population in 1990
reported to have post-primary education (on average, only a third), generate dummy variables
for counties in each quartile of initial education, and include the interactions between these
education quartile dummy variables and year fixed effects. In Panel E, we calculate a Herfind-
ahl index capturing initial concentration in the tradable (primary and secondary) sectors and
include interactions between dummy variables for each quartile of the Herfindahl index and year
32In fact, the estimated coefficients for 2002–2003 are significantly different from the pre-treatment coefficientsfor secondary GDP, although the corresponding tests for exports, GDP, and GDP per capita fail to reject equality.Of the coefficients estimated post–2003, the two cases in which the tests fail to reject compare the coefficientsestimated for exports for 2012–2013 to 2001–2002 (p-value .222) and GDP per capita for 2004–2005 to 2000–2001(p-value .13).
33A small number of observations are missing population data. The results are also consistent if each countyobservation is weighted with respect to the initial total employment or GDP.
22
fixed effects. The results are uniformly consistent.
An additional robustness check explores whether the reform of state-owned enterprises
(SOEs) could be an alternate channel for the observed pattern. In addition to the market
liberalization implemented in this period linked to WTO accession, a major restructuring of
SOEs was implemented starting in the mid-1990s and accelerating in the latter part of the
decade (Naughton, 2007). Unfortunately, no county-level data are available on SOE employ-
ment. However, we can construct a county-level proxy using data on SOE employment in broad
sectors (agriculture, mining and manufacturing) in each province as a percentage of total sector
employment in that province in 1996 (the first year in the sample). We then use the 1990 em-
ployment weights by sector to construct a county-level average.34 Cross-county variation in the
imputed baseline share of SOE employment is thus generated by variation across counties in the
salience of agriculture, mining and manufacturing, and variation across provinces in the relative
importance of SOE employment in these three sectors. We then construct dummy variables for
counties in each quartile of the initial imputed SOE fraction, and interact these dummies with
year fixed effects in the main specification. The results are reported in Panel F of the same
table, and they are entirely consistent with the main specifications.
Finally, we re-estimate the results employing as the dependent variable the average night
lights index within county borders as a proxy for the intensity of local economic activity. This
addresses the potential challenge introduced by measurement error or selective misreporting
in county yearbook data. We observe a high correlation between the night lights index and
reported county-level GDP. When the primary specification is re-estimated using the night
lights index as a dependent variable, the estimated relationship is significant and positive at
the one percent level, and suggests that a county moving from the 25th to the 75th percentile
of the NTR gap shows evidence of a 5% relative increase in night brightness post-2001.
4.4 Firm-level outcomes
The county-level data previously used do not include data on some key outcomes of interest:
particularly, capital investment, foreign investment and wages. In addition, the data on sectoral
employment and exports are very limited. As an additional source of evidence, we utilize the
large-scale industrial survey collected from 1998 to 2008, a data source described in detail in
34These employment data are drawn from the national statistical yearbooks; data on SOE employment inthe highly disaggregated subsectors reported in the census are unavailable until much later, in the post–WTOperiod. Unsurprisingly, SOE employment is close to zero in agriculture (averaging 2%) and near universal inmining (averaging 91%). The SOE share in manufacturing employment is variable, with a mean of 38% and astandard deviation of 13%.
23
Brandt et al. (2012). The data are collected in annual surveys conducted by the National Bureau
of Statistics, and they include all state-owned industrial firms (in mining, manufacturing, and
public utilities) and all non-state firms in the same sectors with sales above 5 million yuan. For
this analysis, we restrict the sample to manufacturing firms.
A variety of firm-level outcomes are observed. Employment and the total wage bill are
directly reported, enabling us to estimate the average wage per worker. The perpetual inventory
method is used to estimate the capital stock, as the firm’s founding year is also reported; the
average growth rate observed at the province-sector level over the sample years is used to
estimate average annual investment rates. We also use the estimate of the capital stock to
calculate firm-level capital intensity (the ratio of the capital stock to total employment), and a
define a dummy variable equal to one if the firm reports any foreign-owned capital. For export
values, sales, value added and exports, we use the deflators constructed by Brandt et al. (2012)
to construct constant-price estimates.
The firms can be geographically linked only to the prefecture, as county indicators are un-
available. Accordingly, we perform this analysis at the prefecture level; the dependent variables
are calculated as the sum of the relevant firm-level variables within the prefecture and year, to
capture the total size of the large-scale manufacturing sector. (For capital intensity, the wage,
and value-added per worker, the mean is employed.) The NTR gap is calculated as the mean
of the NTR gap across all constituent counties in the prefecture and is denoted Yfpt for the
NTR gap in prefecture f and province p. The same control variables are also included and
are calculated as the prefecture-level mean, and prefecture and province-year fixed effects are
included.35
Yfpt = β1Postt ×NTRGapfp + X ′fptθ + γpt + νf + εcfpt (5)
The results are reported in Table 5; again, the coefficients and means reported correspond
to prefecture-level aggregates of the firm data. The first two columns in Panel A show that an
increase in tariff uncertainty from the 25th to the 75th percentile of the prefecture NTR gap
is associated with a 11% increase in large-scale manufacturing employment in the prefecture
post–2001. The real capital stock shows a 10% increase. The mean probability that a firm
reports any foreign capital increases by 10%. Given that the expansions of capital and labor
35In the primary results, we do not construct a weighted mean; however, the results are also consistent if weconstruct means at the prefecture-year level that are weighted with respect to county population. These resultsare reported in Table A13 in the Appendix.
24
are of roughly equal magnitude, there is no significant shift in capital intensity, as reported in
Column (4). Finally, in Columns (5) and (6), we observe that the total wage bill increases by
23%, corresponding to a 2% increase in the average wage per worker.
In Panel B, we report results for additional outcomes: total exports, sales, value added,
value added per worker and profits of the sampled firms at the prefecture level. We observe
that a prefecture moving from the 25th to the 75th percentile of the NTR gap will experience
an increase in exports of 21% and increases in sales, value added and profits of between 25%
and 51%. Value added per worker increases by 3%. These results are generally somewhat larger
than those employing county-level data, suggesting that the effects of reduced tariff uncertainty
may be larger for above-scale firms.36
Moreover, the previous results were estimated only for a subsample of counties reporting
export data, while these data include all prefectures, thus enabling us to verify that the increase
in exports is observed consistently across the larger sample. Similarly, value added per worker
can be calculated only for a small sample in the county-level data, given the limited number of
observations reporting both employment and value added. However, the evidence of an increase
in value added per worker is consistent with the previously reported evidence from county-level
data that an increase in the NTR gap from the 25th to the 75th percentile generates a 5%
increase in secondary employment and a larger, 9% increase in secondary value added. We can
also verify using a different source of data the effects observed in the county-level data, including
a modest increase in secondary employment and substantial increases in exports, sales, profits,
and value added in the secondary sector.
As corroborating evidence for the observed increase in foreign investment here, we also
estimate parallel specifications using a small sample of counties that report FDI data, as well
as provincial-level data on FDI. The results are reported in Table A12 in the Appendix, and
show an increase in FDI at both the county and the provincial level. The county-level data
suggest that a county moving from the 25th to the 75th percentile of the NTR gap would show
an increase of around 40% in contracted FDI.
36In addition, 20% of the firms in this sample are state-owned or collective firms; on average, the level ofexports observed in these firms is approximately one-sixth the level of exports observed for non-state firms. Thereported increase in exports, sales and value added (among other variables) is not observed in the subsample ofSOE firms. Thus the main results seem to be driven by firms outside the state sector.
25
4.5 Mechanisms
Returning to the conceptual framework, it is useful to highlight the mechanisms that generate
the observed patterns of accelerated structural transformation post-WTO accession in counties
more exposed to tariff uncertainty ex ante. First, we observe both a substantial increase in
exports and an increase in foreign direct investment. Both effects are evident in county-level
data, as well as data derived from a survey of large firms.
Second, as previously noted, there is fairly robust evidence of substitution of productive
factors out of agriculture in counties characterized by higher ex ante NTR gaps following WTO
accession. The effects previously reported were of relatively small magnitude; a shift from
the 25th to the 75th percentile of the NTR gap generated a relative decline in agricultural
investment and value added post-WTO of no more than 1-3%. However, the sample does
include some urban counties; approximately 5% of counties reported in the 1990 census that
less than one-fifth of the population was engaged in the primary sector. If we restrict the sample
to exclude these counties, the observed substitution out of the primary sector roughly doubles
in magnitude. In addition, we observe increased investment and output in both the secondary
and tertiary sectors, although the effects are larger in the secondary sector.
The growth of the secondary sector as the primary sector shrinks is consistent with both
the reallocation and the local demand channels. However, the fact that non-tradable (tertiary)
production is expanding suggests that the local demand effect dominates the reallocation effect
for the tradable sector. In addition, we can document that the reduction in tariff uncertainty
seems to generate an increase in returns to factors in the medium-term, as evident in the persis-
tent increase in wages and value added per worker observed in the firm data. This is consistent
with the hypothesis that there are barriers to full mobility of capital and labor that slow the
equalization of factor returns across counties. Alternatively, there may be positive agglomera-
tion effects in export production that lead to persistently more rapid growth in counties that
benefit from the reduction in tariff uncertainty post-WTO.
5 Additional robustness checks
In this section, we present additional robustness checks, including placebo tests that corroborate
the hypothesis that the main effects are driven by reduced uncertainty in the U.S. market;
analysis of cross-country spillovers; and finally, evidence of heterogeneous effects.
26
5.1 Placebo analysis
Throughout this analysis, we have assumed that the discontinuous shock experienced by China
at the point of its WTO accession is a decrease in tariff uncertainty in the U.S. market. Here,
we implement a placebo analysis to evaluate this assumption. As previously noted, the EU
endowed China with permanent NTR status in 1980, long before the latter’s accession to the
WTO, and other trading partners (excluding the U.S.) followed suit. Accordingly, China faced
no tariff uncertainty in non-U.S. markets during the period of interest here.
We conduct two placebo tests. The first uses data from the UNCOMTRADE database
reporting China’s exports to all destinations at the 2-digit product level from 1995 to 2013. We
then estimate a simple regression in which the dependent variable is exports of products p to
destination country d in year t, and the independent variables are a post dummy interacted with
the U.S. NTR gap at the product level and a dummy for the U.S., and the post-NTR interaction
interacted with a dummy for the other four top export destinations (the EU, Japan, Korea, and
Taiwan). The specification also includes controls for the product-specific tariff imposed by each
of the five major destinations on each product, summarized Xpdt, and country-year, product-
year, and country-product fixed effects. Standard errors are clustered at the product level.37
Exppdt = β1NTRpt × USd × Postt + β2NTRpt ×Otherdestind × Postt (6)
+ Xpdt + ωdt + ξpt + µdp + εdpt
We hypothesize that β1 will be positive and significant, and β2 should not be significantly
different from zero: products characterized by a larger NTR gap exhibit a disproportionate
increase in exports to the U.S. post-WTO accession, but there should be no significant increase
in exports to other major destinations. The results are reported in Panel A of Table 6, and we
observe exactly the postulated pattern; in Columns (3) and (4), quadratic controls for tariffs
are also included. β1 is positive and β2, while positive, is insignificant and approximately one-
tenth the magnitude of β1. The final row of the table shows that the hypothesis that the
coefficients are equal in magnitude can be rejected at the one percent level. This evidence
is again consistent with the assertion that the key immediate shock experienced with WTO
accession was a reduction in trade uncertainty in the U.S. market, not a shock in other major
export destinations.
Second, we conduct a placebo test by constructing an artificial “EU gap”, comparing the
37Similar results are observed if standard errors are clustered at the partner level.
27
EU tariff rates imposed on countries subject to relatively high tariffs to the tariff rates imposed
on Chinese goods. Since the EU does not specify a non-NTR set of countries, we simply identify
for each industry represented in the Chinese data the five trading partners on which the EU
imposes the highest tariffs for that industry and calculate a “maximum tariff” that is the mean
of these tariffs. We then calculate a placebo “EU gap” equal to the difference between these
high tariffs and the tariff imposed on Chinese goods, and follow the same procedure previously
utilized to construct a county-level EU gap that varies across counties and over time.
To conduct the placebo analysis, we estimate the following specification, regressing county-
level outcomes on the EU placebo gap, using the same control variables and fixed effects included
in the main specification. We also control flexibly for the EU high tariff rate EUcfpt.38
Ycfpt = β1EU Gapcfpt + EUcfpt + X ′cfptθ + γpt + νf t+ Urbcfp × (γpt + νf t) + δc + εcfpt (7)
The results are reported in Table 6, and the estimated coefficients are small in magnitude,
insignificant and varying in sign. This suggests that there is no evidence that tariff variation
orthogonal to China’s export expansion predicts cross-county variation in economic outcomes.
5.2 Spillovers from one county to another
Thus far, the analysis has ignored any possible spatial spillovers in the positive shocks to export
production following China’s WTO accession. A positive shock to the export sector in one
county may have positive effects on the economies of neighboring counties through several
channels. First, increased income may generate positive demand shocks for goods and services
produced in adjacent counties. Second, productive factors may shift across county lines in
response to more rapid growth. Third, if there are positive agglomeration effects in exporting
industries, growth in secondary exports may directly stimulate exporting in adjacent counties.
We focus on the prefecture as a unit capturing the local region. Our objective is to match
each county to other counties in the same prefecture within certain specified geographic ranges
(0–25 kilometers, 26–50 kilometers, and 51–75 kilometers), where straight lines between the
county centroids are used to calculate these distances. If multiple counties fall within a specified
distance range, the closest county is employed as the match. (Counties in different prefectures
are not included for the purpose of this analysis.)39
38Specifically, we generate a set of dummy variables for each two-percent range in the distribution of the hightariff rate (50 dummy variables in all) and include these variables, as well as their interaction with the postdummy.
39These results are also consistent if estimated focusing on the nearest county within the same province, rather
28
We then estimate a series of specifications, regressing the outcomes of interest on the in-
teraction of the post dummy and the own-county NTR gap and the interaction of post and
the neighboring county gap. The specification of interest can be written as follows, where the
neighboring county gap is denoted NTRNeighcfp. The control variables include the previously
specified variables measured for the main county and for the adjacent county.
Ycfpt = β1Postt ×NTRGapcfp + β2Postt ×NTRNeighcfp (8)
+ X ′cfptθ + γpt + Urbcfp × (γpt) + δc + εcfpt
The results of estimating equation (8) are reported in Table 7. There is some evidence of
positive spillovers in export production and the associated expansion in secondary, tertiary and
total GDP, especially for counties that are located relatively close to one another. However,
the coefficients for the estimated spillovers are uniformly smaller than those for the own-shock
effects, and they generally decrease as the distance from the neighboring county increases. In
the final row of each panel, we report the p-value for a test of the hypothesis β1 = β2, and this
hypothesis is generally rejected, except for the specifications employing export data.
5.3 Heterogeneous effects
We also present some evidence regarding heterogeneous effects, identifying counties concentrated
in industries that should show a more robust response to the reduction of tariff uncertainty.
In particular, we focus on counties concentrated in industries that are more capital-intensive,
counties concentrated in industries that export a higher proportion of their output to the U.S.,
and counties that include preferential export zones.
Heterogeneity with respect to baseline capital intensity The tariff uncertainty faced
by exporting firms in China prior to WTO accession presumably had a more significant effect
on capital utilization vis-a-vis labor utilization, given that capital investments are generally ir-
reversible. While the county-level panel does not include any detailed information about capital
investment that would allow for a direct test of this hypothesis, we can examine heterogeneous
effects with respect to capital intensity of the industries observed in the county at baseline.
Using the same capital intensity variable constructed from the firm-level survey, we calculate
average capital intensity at the industry level and construct a county-level proxy for capital
than within the same prefecture.
29
intensity in the secondary sector using the 1990 employment weights. (Information about capital
intensity in the primary sector is not available, and thus it is excluded from this analysis.) We
standardize this variable to have mean zero and standard deviation one, and interact it with
the post-NTR gap interaction in the main specification of interest, equation (3).
The results are reported in Panel A of Table 8. We observe that the β1 coefficients show the
familiar pattern. In addition, the interaction terms are positive and significant for exports and
total GDP. This suggests that, consistent with theoretical predictions, a one standard deviation
increase in ex ante estimated capital intensity in the secondary sector at the county level yields
an increase in GDP post-WTO accession that is around 30% larger in relative terms.
Heterogeneity with respect to the U.S. share of exports Unfortunately, the export
data available at the county level do not report the destination of these exports. Accordingly,
we have used the NTR gap as a proxy for tariff uncertainty in the U.S. market without taking
into account how important that tariff uncertainty is for a particular industry.
However, we can use the available UNCOMTRADE data on Chinese exports at the product-
destination level to calculate the proportion of exports destined for the U.S. by industry in 1996.
We then generate a county-specific weighted average, construct a dummy variable for a county
characterized by a U.S. export share above the median, and interact this dummy with the
post-NTR interaction in our main specification. The intuition is that counties concentrated in
industries characterized by a high NTR gap that nonetheless export a significant fraction of
output to the U.S. will exhibit the greatest degree of export-driven expansion post-2001.
The results are reported in Panel B of Table 8. Again, we observe that the main effects
are consistent with the previously estimated coefficients. In addition, the interaction terms
including the high U.S. export dummy variables are also generally significant and positive (with
the exception of per capita income), suggesting that as expected, the most meaningful benefits
of WTO accession are experienced by counties and industries that are both facing significant
tariff uncertainty and export disproportionately to the U.S.
Heterogeneity with respect to export processing zones Throughout the reform period,
China has established various geographically defined zones that offer benefits for foreign and/or
exporting firms. The earliest of these were the five Special Economic Zones established in
Hainan, Guangdong, and Fujian between 1980 and 1984. More recently, dedicated export pro-
cessing zones have been established in 26 urban areas where overwhelmingly foreign enterprises
30
benefit from a range of preferential policies, including rebates of value-added taxes and duties
imposed on any imported components. The EPZs were first established in 2001 and continued
to expand in subsequent years, though enterprises in official export zones still account for a
relatively small fraction of overall exports (Wang and Wei, 2010).
Given that the EPZs focus exclusively on exporting, and particularly on processing exports
that are targeted to higher-income markets like the U.S., it is plausible to assume that there
might be particularly large effects of WTO accession for these areas. To evaluate heterogeneous
effects along this dimension, we re-estimate our primary specification including an interaction
with an export zone dummy, as well as the zone dummy as a linear control.40 The results are
reported in Panel C of Table 8. We can observe that in general the effects of the reduction in
tariff uncertainty are larger for areas including an EPZ, though the difference is significant only
for exports and secondary production.41 However, the positive effects are clearly not restricted
to these zones and remain substantial in the broader sample.
5.4 Aggregate productivity and growth
Finally, it may be useful to present some simple back-of-the-envelope calculations that quantify
the contribution of the reduction in trade uncertainty generated by WTO accession to shifts in
aggregate productivity and growth in China over this period. First, we can quantify the con-
tribution of labor reallocation across sectors (from agricultural production to non-agricultural
production) to aggregate productivity, following McCaig and Pavcnik (2014b). A growing lit-
erature has documented that value added per worker is significantly higher in non-agricultural
compared to agricultural production in developing countries (Gollin et al., 2014), and we can
replicate this stylized fact using the county-level data employed here; value added per worker in
the secondary sector is approximately 6.5 times value added per worker in the primary sector.
Given that the results reported in Table 2 suggest that around 5% of the agricultural labor force
shifted into non-agricultural production following WTO accession, this suggests an increase of
nearly 30% in aggregate productivity driven by this reallocation alone.42
We can also explore the importance of WTO accession in overall growth in county-level
40We follow the list of export zones provided in Table 2A.2 in Wang and Wei (2010), and identify the countiesin our sample that share the same prefecture-city code provided in that source as counties that are part of theexport zone. This does not necessarily mean that the entire area of the county falls within the zone, but there issome overlap.
41There is also a negative interaction term observed for primary GDP.42This calculation ignores any possible labor reallocation into the tertiary sector, for two reasons. First, there
was no significant evidence of an increase in tertiary sector employment in the primary results, although theseresults may be limited by the small sample. Second, it is not possible to estimate value added per worker in thetertiary sector, as value added in this sector is not generally reported.
31
GDP during this period; more details about how these calibrations are conducted can be found
in Appendix A1.2. The average county in this sample shows 227% growth in county-level
GDP from 2002 to 2013 (i.e., in the post-WTO period). Our results suggest that for a county
characterized by an NTR gap at the median prior to WTO accession, the reduction in tariff
uncertainty in the U.S. market results in an increase in GDP of 32%. Accordingly, export-driven
growth enhanced by WTO membership accounts for approximately 15% of overall GDP growth.
(A similar calculation for secondary GDP suggests that growth driven by the WTO accession
shock accounts for approximately 25% of overall secondary growth from 2002 to 2010, the final
year in which secondary GDP is observed for a substantial sample.)
Turning to employment, the average county in our sample that reports detailed employment
data exhibits growth in secondary employment of 78% in total from 2002 to 2010; about 11
percentage points of this growth is attributable to the export shock captured here.43 Using data
on average secondary employment at baseline in 2002, we can conservatively estimate that the
number of additional positions created in the secondary sector as a result of export expansion
to the U.S. is roughly 9.6 million in the sampled counties. According to the 2000 census, the
counties included in the sample account for 75% of total secondary employment in China in that
year; under the assumption that the effects of the NTR shock are identical in the counties not
represented in our sample, we estimate that total secondary employment created equals 12.9
million nationwide. This compares to an estimate of 1.5 million manufacturing positions lost in
the U.S. as a result of Chinese export expansion, as reported in Autor et al. (2013). Thus we
can conclude that while WTO accession is certainly not the only phenomenon generating rapid
economic expansion in China during this period, its importance is non-trivial.
6 Conclusion
In this paper, we use a new panel of county-level data to present the first evidence of the effect
of China’s accession to the WTO in 2001—a policy shift that removed uncertainty over the tariff
rates that Chinese exporters would face in the U.S. market—on structural transformation and
growth. The identification strategy exploits variation across industries in the size of the gap
between the MFN tariffs and the higher tariffs that Chinese producers risked exposure to prior to
WTO accession, as well as variation across counties in the baseline composition of employment in
the secondary sector. We then evaluate whether counties with a high concentration of industries
43Again, 2010 is the last year in which a substantial number of counties report data on secondary employment.More details about these calculations are reported in Appendix A1.2.
32
characterized by large tariff gaps show more rapid growth post–2001.
Our results suggest that counties that benefited most from the reduced tariff uncertainty
show substantial expansion post–2001. Employment, GDP, and value added in the secondary
sector all increase, while the agricultural sector contracts. We also observe a substantial increase
in GDP per capita. Moreover, these patterns are observed only after WTO accession, suggesting
that they do reflect the hypothesized channel of reduced tariff uncertainty, and are not evidence
of ex ante differences in observable characteristics comparing across counties with larger and
smaller NTR gaps.
This paper is the first to present evidence on the impact of the reduction in tariff uncertainty
on structural transformation at the local level in China, and joins a relatively small literature
analyzing the effects of enhanced trade access in stimulating growth in developing countries.
These results highlight the importance of securing access to developed country markets for
developing countries that pursue export-driven growth strategies. Understanding the implica-
tions of U.S. trade for Chinese growth may contribute to a more complete understanding of the
global impact of rising U.S.–China bilateral trade and China’s rise as a global manufacturing
powerhouse over the past two decades.
33
7 Figures and Tables
Figure 1: Variation in Tariff Policy Over Time
(a) China’s Import Tariffs Over Time
(b) Major Trading Partners’ Tariffs Over Time
Notes: The first subfigure shows China’s average domestic import tariff, calculated as the weighted average ofindustry-level tariffs and utilizing as weights the share of total Chinese imports constituted by each industry’s imports.The second subfigure shows the mean tariff imposed on Chinese exports by major trading partners from 1996 to 2013.For each trading partner, we again calculate the weighted average of industry-level tariffs, utilizing as weights the share oftotal Chinese exports constituted by each industry’s exports. Tariff data is obtained from the WITS-TRAINS database.
34
Figure 2: Composition of Employment and GDP in China
(a) Employment
(b) GDP
Notes: This graph presents aggregate statistics for China as a whole from 1990 to 2015, employing data from theNational Bureau of Statistics. The primary sector includes agriculture, forestry and fishing, the secondary sector includesmanufacturing and mining, and the tertiary sector includes services. GDP is reported in billions of constant 2000 yuan.Employment is reported in millions of persons.
35
Figure 3: Bilateral Trade Flows and China’s Import Penetration in the U.S.
Notes: The bilateral trade data is drawn from the IMF’s Direction of Trade Database. Import and exports are deflated to2009 U.S. dollars using the PCE price index. China’s import penetration in the U.S. market is defined as U.S. importsfrom China divided by total US expenditure on goods, measured as U.S. gross output plus U.S. imports minus U.S.exports. The latter three series are drawn from the World Development Indicators.
36
Figure 4: NTR Gap at the County Level
Notes: The figure is a histogram of the gap between normal trade relations (NTR) tariffs and non-NTR tariffs, calculatedat the county level utilizing industry employment shares as reported in the 1990 census as weights.
37
Figure 5: Estimated Dif-in-dif Coefficients and 90% Confidence Intervals
(a) Exports (b) Secondary
(c) GDP (d) GDP per capita
Notes: These graphs report the coefficients on the interaction of dummy variables for each two-year interval and thecounty-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standard deviationone. The specification also includes a number of control variables: an interaction of the post-reform indicator variable anda time-invariant dummy capturing whether the county is characterized by high contract intensity industries, theindustry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate for imports ofcounty-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentage of firmslicensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level.
38
Table 1: Summary Statistics
Variable Mean St. dev. Min. Max. Obs. Num. counties
Total population 509.04 463.9 37 6850.02 27802 1602Total emp. 250.00 164.24 8.20 1544.40 19210 1419Primary employment share .56 .17 .01 .95 2906 345Second employment share .20 .13 0 .71 2906 345Tertiary employment share .24 .09 .04 .83 2906 345Exports 752.81 1915.17 .07 23653.74 5105 1005GDP 8332.22 33312.19 83.16 596015.48 28355 1655GDP per capita 9666.65 13146.19 821.19 239396.95 26316 1595Primary GDP share .29 .16 0 .98 12262 1432Secondary GDP share .39 .15 .02 1 12430 1430Tertiary GDP share .32 .09 .03 .91 12340 1430Sown area 61.54 53.88 0 942.78 8019 966Grain production 230.22 202.63 .17 2581.5 27184 1600Cash production 42.34 44.33 .02 391.43 25779 1559
Notes: This table reports the mean, standard deviation, minimum, maximum, number of observations and number ofcounties reporting any observations for key variables. Total population and employment is reported in thousands ofperson; the employment shares report the percentage of total employment constituted by employment in the specifiedsector. Exports and GDP are reported in millions of yuan and GDP per capita in yuan, deflated to 2000 constant prices;the GDP shares report the percentage of total GDP constituted by the specified sector. Sown area and grain area arereported in thousands of hectares, and grain production and cash crop production are reported in thousands of tons.
39
Table 2: Primary Results
(1) (2) (3) (4) (5) (6)
Panel A: Exports and GDP
Exports Primary Secondary Tertiary GDP Per capita
Post x NTR gap 248.261 42.915 4613.042 1875.907 2655.506 955.651(131.466)∗ (32.992) (1254.208)∗∗∗ (870.109)∗∗ (744.508)∗∗∗ (400.357)∗∗
Mean dep. var. 752.808 1010.083 6371.597 4642.585 8332.215 9666.646Counties 1005 1465 1458 1455 1655 1595Obs. 5105 13034 13053 12922 28355 26316
Panel B: Employment
Primary Secondary Tertiary Agri. Total emp. Total pop.
Post x NTR gap -5.006 7.678 -1.327 -7.860 -.557 35.040(6.941) (4.784) (3.122) (4.220)∗ (2.713) (12.241)∗∗∗
Mean dep. var. 173.968 66.366 72.675 217.098 250.002 509.042Counties 353 398 399 1305 1419 1602Obs. 3127 3376 3501 20634 19210 27802
Panel C: Agriculture and value added
Sown area Agri. Grain Cash Primary Secondarymachine value added value added
Post x NTR gap .152 -3.068 -8.078 -1.400 -2.615 51.358(.964) (1.585)∗ (4.772)∗ (1.007) (1.463)∗ (19.680)∗∗∗
Mean dep. var. 61.54 33.838 230.222 42.341 86.375 223.996Counties 966 1621 1600 1559 1574 1560Obs. 8019 27277 27184 25779 27054 27126
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level.
In Panel A, the dependent variables include exports at the county level; primary, secondary, tertiary and total GDP; andper capita GDP. Exports and GDP are reported in millions of yuan deflated to 2000 constant prices; per capita GDP isreported in yuan, similarly deflated. In Panel B, the dependent variables include employment in the primary, secondaryand tertiary sectors, total employment, and population, all reported in thousands of persons. In Panel C, the dependentvariables include sown area reported in thousands of hectares, agricultural machinery reported in 10,000 kilowatts, grainand cash crop production reported in thousands of tons, and primary and secondary value added reported in millions ofyuan deflated to 2000 constant prices. Asterisks indicate significance at the ten, five and one percent level.
40
Table 3: Robustness Checks
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: NTR gaps estimated using 2000 employment weights
Post x NTR gap 318.819 50.079 2347.465 839.859 1702.840 1427.298(148.609)∗∗ (32.792) (543.962)∗∗∗ (445.581)∗ (394.809)∗∗∗ (288.117)∗∗∗
Obs 5130 13387 13406 13277 28769 26584
Panel B: NTR gaps estimated assigning non-tradables zero weights
Post x NTR gap 207.145 20.648 2901.603 1364.624 834.090 541.618(94.166)∗∗ (27.721) (733.017)∗∗∗ (507.601)∗∗∗ (647.102) (229.470)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel C: Main specification controlling for the share of non-primary employment
Post x NTR gap 340.394 49.207 4899.748 2022.489 3021.323 1011.088(154.676)∗∗ (34.178) (1332.319)∗∗∗ (920.674)∗∗ (782.134)∗∗∗ (427.384)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel D: Main specification controlling for the share of high gap employment
Post x NTR gap 662.396 33.080 5322.362 2259.664 3075.183 1478.209(241.783)∗∗∗ (35.298) (1607.042)∗∗∗ (1137.114)∗∗ (932.773)∗∗∗ (485.119)∗∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel E: Heterogeneity with respect to import competition
Post x NTR gap 259.677 47.008 4421.264 1864.949 2674.020 946.560(137.453)∗ (30.492) (1178.399)∗∗∗ (817.705)∗∗ (726.542)∗∗∗ (388.232)∗∗
High import int. -85.652 13.529 -1011.705 -119.101 -14.467 -400.319(73.675) (21.190) (488.629)∗∗ (362.105) (468.540) (256.720)
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level. The dependent variables are exports at the county level; primary,secondary, tertiary and total GDP; and per capita GDP.
In Panel A, the NTR gap at the county level is estimated using employment weights from the 2000 census. In Panel B,the NTR gap is estimated using employment weights from the 1990 census and assigning the services or non-tradablesector a zero weight. In Panel C, the specification includes a full set of interactions between dummies for quartiles ofinitial secondary and tertiary employment as a fraction of total employment and year fixed effects. In Panel D, thespecification includes a full set of interactions between dummies for quartiles of employment in each of five high NTR gapindustries as a fraction of total employment and year fixed effects. In Panel E, the post-NTR interaction is interactedwith a dummy equal to one for districts who are above the median of soybeans and cotton as a fraction of total sownarea, and we also include interactions between quartiles of this fraction and year fixed effects. Asterisks indicatesignificance at the ten, five and one percent level.
41
Table 4: Estimating the Effect of the NTR Gap over Time
Exports Secondary GDP Per capita(1) (2) (3) (4)
NTR gap x 97-98 -555.328 -468.512 -96.020(889.739) (816.436) (342.157)
NTR gap x 99-01 169.953 -829.094 -1913.869 492.171(266.857) (1208.239) (1336.734) (569.309)
NTR gap x 02-03 280.018 2379.107 -1112.166 755.269(309.379) (1225.792)∗ (1951.345) (621.685)
NTR gap x 04-05 1606.207 8406.023 4457.992 2061.496(645.787)∗∗ (2268.483)∗∗∗ (1007.391)∗∗∗ (801.503)∗∗
NTR gap x 06-07 1902.506 11304.270 8062.079 3800.560(618.678)∗∗∗ (2697.777)∗∗∗ (1443.048)∗∗∗ (849.838)∗∗∗
NTR gap x 08-09 2066.532 13786.760 11521.510 3812.086(614.017)∗∗∗ (3515.696)∗∗∗ (2255.627)∗∗∗ (1121.989)∗∗∗
NTR gap x 10-11 2344.534 16556.580 11302.720 5315.567(759.210)∗∗∗ (3906.195)∗∗∗ (2084.389)∗∗∗ (1109.795)∗∗∗
NTR gap x 12-13 1749.813 18445.150 9108.340 6998.901(1069.573) (3924.502)∗∗∗ (1619.350)∗∗∗ (1235.621)∗∗∗
Obs. 5105 13053 28355 26316
Notes: The independent variables are the interaction of dummy variables for each two-year interval and the county-levelgap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standard deviation one. Thespecification also includes a number of control variables: an interaction of the post-reform indicator variable and atime-invariant dummy capturing whether the county is characterized by high contract intensity industries, theindustry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate for imports ofcounty-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentage of firmslicensed to export, and the industry-weighted time-varying NTR rate). Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level. The dependent variables are exports at the county level; primary,secondary, tertiary and total GDP; and per capita GDP. Asterisks indicate significance at the ten, five and one percentlevel.
42
Table 5: Factor utilization and other firm outcomes
(1) (2) (3) (4) (5) (6)
Panel A: Factor utilization
Emp. Capital Foreign capital Cap. Wages Wages perdummy intensity worker
Post x NTR gap 20.010 591.271 .013 -1.154 507.123 .258(9.469)∗∗ (229.141)∗∗∗ (.006)∗∗ (1.780) (191.224)∗∗∗ (.135)∗
Mean dep. var. 116.35 3655.667 .077 61.848 1354.195 9.416Obs. 2403 2004 2426 1981 2426 2403
Panel B: Other firm outcomes
Exports Sales Value added Profits VA per worker
Post x NTR gap 1128.778 10290.790 2589.090 1776.065 2.891(570.918)∗∗ (3170.815)∗∗∗ (688.696)∗∗∗ (544.468)∗∗∗ (1.269)∗∗
Mean dep. var. 3354.639 24658.011 5212.273 2143.398 51.59Obs. 2426 2426 2213 2426 2189
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the prefecture-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification includes the same control variables described in the notes to Table 2, all calculated asthe average at the prefecture-year level, as well as province-year and prefecture fixed effects. All standard errors areestimated employing clustering at the prefecture level.
The dependent variables in Column (1) include total employment in sampled firms, the total wage bill in sampled firms,mean wage per worker, total capital stock in sampled firms, and mean capital intensity. The dependent variables inColumn (2) include total exports, sales, value added and profits in sampled firms, as well as mean value added perworker. Asterisks indicate significance at the ten, five and one percent level.
43
Table 6: Placebo Tests
(1) (2) (3) (4) (5) (6)
Panel A: Chinese exports to all destinations
ExportsPost x NTR Gap x U.S. 8.891 8.893
(.303)∗∗∗ (.303)∗∗∗
Post x NTR Gap x Other four 1.936 1.946(1.407) (1.408)
Test β1 = β2 .000 .000Obs. 866857 866857
Panel B: EU NTR Gap
Exports Primary Secondary Tertiary GDP Per capita
EU NTR gap -145.020 1.603 545.063 -263.574 293.795 222.871(207.788) (15.931) (415.354) (625.505) (259.916) (145.574)
Obs. 5105 13034 13053 12922 28355 26316
Notes: In Panel A, the independent variables are the U.S. NTR gap at the product level interacted with post interactedwith a dummy for the U.S., and the post-NTR interaction interacted with a dummy for the other four top exportdestinations (the EU, Japan, Korea, and Taiwan). The dependent variable is China’s exports to all destinations at the2-digit product level from 1995 to 2013, as reported in the UNCOMTRADE database. The specification also includescontrols for the product-specific tariff imposed by each of the five major destinations on each product, and country-year,product-year, and country-product fixed effects; in the second specification, quadratic controls for product-specific tariffsare added. Standard errors are clustered at the product level.
In Panel B, the independent variable is the interaction of a dummy variable equal to one for the post–2001 period and thegap between the highest EU tariff observed for a given industry and the tariff imposed on Chinese exports of thatindustry’s output to the EU, standardized to have mean zero and standard deviation one. The specification also includesa number of control variables: an interaction of the post-reform indicator variable and a time-invariant dummy capturingwhether the county is characterized by high contract intensity industries, the industry-weighted MFA quota fill rate forcounty-produced goods, the industry-weighted national tariff rate for imports of county-produced goods and the tariffrate interacted with the post dummy, the industry-weighted percentage of firms licensed to export, and theindustry-weighted time-varying NTR rate). Province-year and county fixed effects and prefecture-specific trends areincluded, and the time-varying trends are interacted with an urban dummy. We also control flexibly for the high EUtariff (the average tariff imposed on the five trading partners characterized by the highest tariff rates) by constructingfifty dummy variables corresponding to different percentiles of the EU tariff distribution, generating dummy variable fixedeffects, and also interacting those fixed effects with the post dummy. Standard errors are estimated employing clusteringat the county level. The dependent variables are exports at the county level; primary, secondary, tertiary and total GDP;and per capita GDP. Asterisks indicate significance at the ten, five and one percent level.
44
Table 7: Estimating Effects of Spillovers
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: Neighboring counties within 25 kilometers
Post x NTR gap 599.028 -38.835 1154.935 538.699 1587.102 3108.549(274.605)∗∗ (28.203) (258.994)∗∗∗ (171.498)∗∗∗ (619.140)∗∗ (987.168)∗∗∗
Post x NTR gap neighbor 228.002 34.798 388.286 122.400 440.122 1080.828(132.373)∗ (17.814)∗ (180.108)∗∗ (106.263) (260.716)∗ (365.186)∗∗∗
Test β1 = β2 .180 .075 .014 .029 .083 .086Obs. 3415 8298 8309 8272 19258 18612
Panel B: Neighboring counties within 50 kilometers
Post x NTR gap 639.604 -31.570 1254.132 630.973 2206.660 3702.587(362.001)∗ (22.901) (284.955)∗∗∗ (165.196)∗∗∗ (700.847)∗∗∗ (947.026)∗∗∗
Post x NTR gap neighbor 142.073 36.683 589.542 263.226 652.256 244.298(126.646) (13.986)∗∗∗ (286.909)∗∗ (152.923)∗ (350.236)∗ (332.926)
Test β1 = β2 .239 .035 .065 .083 .034 .001Obs. 3703 8693 8710 8687 19665 18947
Panel C: Neighboring counties within 75 kilometers
Post x NTR gap 1040.096 -49.021 1324.951 619.096 2449.197 3716.560(282.241)∗∗∗ (25.894)∗ (343.528)∗∗∗ (217.150)∗∗∗ (745.283)∗∗∗ (1050.409)∗∗∗
Post x NTR gap neighbor 105.155 10.967 113.833 73.402 67.499 -398.609(131.588) (20.421) (199.041) (115.858) (296.505) (381.389)
Test β1 = β2 .006 .141 .001 .02 .003 0Obs. 3110 7264 7282 7254 17526 16938
Notes: The primary independent variables are the interaction of a dummy variable equal to one for the post–2001 periodand the gap between NTR tariffs and the non-NTR rate, standardized to have mean zero and standard deviation one; andthe same post-NTR interaction for a neighboring county. The neighboring county is identified as the closest neighboringcounty within the same prefecture in the specified distance range (0–25 kilometers, 26–50 kilometers, and 51–75kilometers), where distance is calculated as the straight-line distance between the county centroids. The specification alsoincludes a number of control variables: an interaction of the post-reform indicator variable and a time-invariant dummycapturing whether the county is characterized by high contract intensity industries, the industry-weighted MFA quota fillrate for county-produced goods, the industry-weighted national tariff rate for imports of county-produced goods and thetariff rate interacted with the post dummy, the industry-weighted percentage of firms licensed to export, and theindustry-weighted time-varying NTR rate). Province-year and county fixed effects and prefecture-specific trends areincluded, and the time-varying trends are interacted with an urban dummy. Standard errors are estimated employingclustering at the county level. The dependent variables are exports at the county level; primary, secondary, tertiary andtotal GDP; and per capita GDP. Asterisks indicate significance at the ten, five and one percent level.
45
Table 8: Heterogeneous Effects
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: Heterogeneity with respect to baseline capital intensity
Post x NTR gap 230.190 48.796 4521.833 1698.167 2539.418 953.746(127.969)∗ (32.618) (1264.955)∗∗∗ (901.822)∗ (697.220)∗∗∗ (400.407)∗∗
Post x NTR gap x 194.301 -24.197 331.853 656.337 719.340 -49.289Baseline capital intensity (115.922)∗ (21.267) (400.033) (433.018) (331.275)∗∗ (177.073)
Obs. 5105 13034 13053 12922 28355 26316
Panel B: Heterogeneity with respect to the proportion of exports directed to the U.S.
Post x NTR gap 195.729 -39.874 1534.967 464.867 1347.789 2533.753(104.671)∗ (43.884) (548.618)∗∗∗ (379.103) (601.449)∗∗ (544.889)∗∗∗
Post x NTR gap x 60.885 86.813 3229.706 1486.869 1479.826 -1903.951High US exports (126.330) (45.977)∗ (976.256)∗∗∗ (676.696)∗∗ (568.501)∗∗∗ (564.000)∗∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel C: Heterogeneity with respect to export zones
Post x NTR gap 133.152 83.475 2934.297 1483.208 2308.695 881.882(100.254) (27.864)∗∗∗ (975.610)∗∗∗ (617.541)∗∗ (675.818)∗∗∗ (386.779)∗∗
Post x NTR gap x 835.802 -132.990 5261.475 1616.162 3021.898 539.058Export zone (455.256)∗ (50.643)∗∗∗ (1110.446)∗∗∗ (1005.991) (3214.047) (941.209)
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the gap between NTR tariffs and the non-NTR rate, standardized to have mean zero and standard deviation one.The control variables and fixed effects included are identical to those reported in the notes to Table 2. Standard errorsare estimated employing clustering at the county level. The dependent variables include exports at the county level;primary, secondary, tertiary and total GDP; and per capita GDP.
In Panel A, an interaction with estimated capital intensity calculated in 1998 at the county-level is included. In Panel B,an interaction with a dummy variable equal to one if the estimated proportion of exports directed to the U.S., ascalculated using the UNCOMTRADE data from 1996, is above the median is included. In Panel C, an interaction with adummy for export zones is included, and the export zone dummy also enters linearly given that it is time-varying.Asterisks indicate significance at the ten, five and one percent level.
46
References
Acemoglu, Daron, David Autor, David Dorn, Gordon Hanson, and Brendan Price,“Import competition and the great U.S. employment sag of the 2000s,” Journal of LaborEconomics, 2016, 34 (S1), S141–S198.
Anukriti, S and Todd J. Kumler, “Tariffs, Social Status, and Gender in India,” IZA Dis-cussion Papers 7969, Institute for the Study of Labor (IZA) 2014.
Autor, David, David Dorn, and Gordon Hanson, “The China syndrome: The impactof import competition on U.S. labor markets,” American Economic Review, 2013, 103 (6),2121–2168.
, David Zorn, and Gordon Hanson, “The China shock: Learning from labor marketadjustment to large changes in trade,” Annual Review of Economics, 2016, 8, 205240.
Bai, Xue, Kala Krishna, and Hong Ma, “How you export matters: Export mode, learningand productivity in China,” Journal of International Economics, 2017, 104, 122–137.
Balsvik, Ragnhild, Sissel Jensen, and Kjell G. Salvanes, “Made in China, sold in Norway:Local labor market effects of an import shock,” Journal of Public Economics, 2015, 127 (10),137–144.
Bernanke, Ben S, “Irreversibility, uncertainty, and cyclical investment,” The Quarterly Jour-nal of Economics, 1983, 98 (1), 85–106.
Bhattasali, Deepak, Shantong Li, and Will Martin, “Impacts and policy implications ofWTO accession for China,” in Deepak Bhattasali, Shantong Li, and Will Martin, eds., Chinaand the WTO: Accession, policy reform and poverty reduction strategies, Washington, D.C.:World Bank, 2004, pp. 1–16.
Bloom, Nick, Stephen Bond, and John Van Reenen, “Uncertainty and investment dy-namics,” The review of economic studies, 2007, 74 (2), 391–415.
Brandt, Loren and Peter Morrow, “Tariffs and the organization of trade in China.,” 2014.Forthcoming, Journal of International Economics.
, Johannes Van Biesebroeck, and Yifan Zhang, “WTO accession and performance ofChinese manufacturing,” 2015.
, , Luhang Wang, and Yifan Zhang, “Creative accounting or creative destruction?Firm-level productivity growth in Chinese manufacturing,” Journal of Development Eco-nomics, 2012, 97 (2), 339–351.
, Trevor Tombe, and Xiaodong Zhu, “Factor Market Distortions Across Time, Spaceand Sectors in China,” Journal of Economic Dynamics, 2013, 16 (1), 3958.
Bustos, Paula, Bruno Caprettini, and Jacopo Ponticelli, “Agricultural productivity andstructural transformation: Evidence from Brazil,” 2015.
Chen, Ting and J. K.-S. Kung, “Do land revenue windfalls create a political resource curse?Evidence from China,” Journal of Development Economics, 2016, 123, 86–106.
Cheng, Leonard and Yum Kwan, “What are the determinants of the location of foreigndirect investment? The Chinese experience,” Journal of International Ecoomics, 2000, 51,379–400.
47
Cheong, Ching and Ching Hung Yee, Handbook on China’s WTO Accession and Its Im-pacts, Hong Kong: World Scientific, 2003.
Chiquiar, Daniel, “Globalization, regional wage differentials and the StolperSamuelson The-orem: Evidence from Mexico,” Journal of International Economics, 2008, 74, 7093.
Costa, Francisco, Jason Garred, and Joo Paulo Pessoa, “Winners and losers from acommodities-for-manufactures trade boom,” Journal of International Economics, 2016, 102,5069.
Dauth, Wolfgang, Sebastian Findeisen, and Jens Jens Sudekum, “The Rise of the eastand the far east: German labor markets and trade integration,” Journal of the EuropeanEconomic Association, 2014, 12 (6), 1643–1675.
, , and Jens Sudekum, “Trade and manufacturing jobs in Germany,” 2017. Forthcoming,American Economic Review Papers & Proceedings.
Dix-Carneiro, Rafael and Brian K. Kovak, “Trade reform and regional dynamics: Evi-dence from 25 years of Brazilian matched employer-employee data,” 2015.
Dixit, Avinash, “Entry and exit decisions under uncertainty,” Journal of political Economy,1989, 97 (3), 620–638.
Donoso, Vicente, Victor Martin, and Asier Minondo, “Does competition from Chinaraise the probability of becoming unemployed?,” Social Indicators Research, 2015, 130, 373–394.
Edmonds, Eric, Nina Pavcnik, and Petia Topalova, “Trade Adjustment and HumanCapital Investments: Evidence from Indian Tariff Reform,” American Economic Journal:Applied Economics, 2010, 2 (4), 42–75.
Facchini, Giovanni, Anna Maria Mayda, and Minghai Zhou, “The impact of China’sWTO accession on internal migration,” 2016.
Feng, Ling, Zhiyuan Li, and Deborah Swenson, “Trade Policy Uncertainty and Exports:Evidence from China’s WTO Accession,” NBER Working Papers 21985, National Bureau ofEconomic Research, Inc 2016.
Foster, Andrew and Mark Rosenzweig, “Agricultural productivity growth, rural economicdiversity, and economic reforms: India, 1970–2000,” Economic Development and CulturalChange, 2004, 52 (3), 509–542.
Gollin, Douglas, David Lagakos, and Michael Waugh, “The agricultural productivitygap,” Quarterly Journal of Economics, 2014, 129 (2), 939–993.
Handley, Kyle and Nuno Limao, “Trade and investment under policy uncertainty: theoryand firm evidence,” American Economic Journal: Economic Policy, 2015, 7 (4), 189–222.
and Nuno Limao, “Policy uncertainty, trade and welfare: Theory and evidence for Chinaand the U.S.,” 2017. Forthcoming, American Economic Review.
Hasan, Rana, Devashish Mitra, and Beyza Ural Marchand, “Trade Liberalization,Labor-Market Institutions, and Poverty Reduction: Evidence from Indian States,” IndiaPolicy Forum, 2006, 3 (1), 71–122.
48
, , Priya Ranjan, and Reshad Ahsan, “Trade liberalization and unemployment: Theoryand evidence from India,” Journal of Development Economics, 2012, 97 (2), 269–280.
Hornbeck, Richard and Pinar Keskin, “Does agriculture generate local economicspillovers? Short-run and long-run evidence from the Ogallala Aquifer,” American EconomicJournal: Economic Policy, 2015, 7, 192–213.
Jiangxi Paper Industry Co. Ltd., “Four Changes in the Macro-Environment,” CompanyPress Release 2000. July 21, 2000.
Khandelwal, Amit, Peter K. Schott, and Shang-Jin Wei, “Trade liberalization and em-bedded institutional reform: Evidence from Chinese exporters,” American Economic Review,2013, 103 (6), 2169–2195.
, Peter Schott, and Shang-Jin Wei, “Trade Liberalization and Embedded InstitutionalReform: Evidence from Chinese Exporters,” American Economic Review, 2013, 103 (6),2169–95.
Kovak, Brian K., “Regional Effects of Trade Reform: What is the Correct Measure of Liber-alization?,” American Economic Review, 2013, 103 (5), 1960–1976.
Lagakos, David and Michael Waugh, “Selection, agriculture, and cross-country productiv-ity differences,” American Economic Review, 2013, 103 (2), 948–80.
Leight, Jessica, “Complementarity between non-agricultural and agricultural shocks in localindustrialization: Evidence from China,” 2016.
Liu, Qing and Hong Ma, “Trade Policy Uncertainty and Innovation: Firm Level Evidencefrom China’s WTO Accession,” 2016.
Long, Guoqiang, “China’s policies on FDI: Review and evaluation,” in Theodore Moran,Edward M. Graham, and Magnus Blomstrom, eds., Does foreign direct investment promotedevelopment?, Washington, D.C.: Peterson Institute Press, 2005, pp. 315–336.
Manova, Kalina and Zhiwei Zhang, “Multi-product firms and product quality,” 2012.NBER Working Paper 18637.
Marden, Samuel, “The agricultural roots of industrial development: ‘Forward linkages’ inreform era China,” 2015.
McCaig, Brian, “Exporting out of poverty: Provincial poverty in Vietnam and U.S. marketaccess,” Journal of International Economics, September 2011, 85 (1), 102–113.
, “Exporting out of poverty: Provincial poverty in Vietnam and U.S. market access,” Journalof International Economics, 2011, 85 (1), 102–113.
and Nina Pavcnik, “Export markets and household business performance: Evidence fromVietnam,” 2014.
and , “Export markets and labor allocation in a low-income country,” 2014. NBERWorking Paper 20455.
, Margaret McMillan, and Inigo Verduzgo Gallo, “Trade liberalization and labormarket adjustment in Botswana,” 2017.
49
McMillan, Margaret, Dani Rodrik, and Inigo Verduzco-Gallo, “Globalization, struc-tural change, and productivity growth, with an update on Africa,” World Development, 2014,63, 1132.
Naughton, Barry, The Chinese Economy: Transitions and Growth, MIT Press, 2007.
Nunn, Nathan, “Relationship-Specificity, Incomplete Contracts, and the Pattern of Trade,”The Quarterly Journal of Economics, 2007, 122 (2), 569–600.
Pierce, Justin and Peter Schott, “Does trade liberalization with China influence U.S.elections?,” 2016. NBER Working Paper 22178.
and , “The surprisingly swift decline of U.S. manufacturing employment,” AmericanEconomic Review, 2016, 106 (7), 1632–1662.
and , “Trade liberalization and mortality: Evidence from U.S. counties,” 2016.
Rowley, Storer H., “China Woos Western Businesses, Snubs Clinton,” Chicago Tribune 1993.May 20, 1993.
Rumbaugh, Thomas and Nicolas Blancher, “China: International trade and WTO acces-sion,” 2004. IMF Working Paper.
Shanghai Securities News, “Outlook 2000,” Shanghai Securities News 1999. December 28,1999.
South China Morning Post, “HK business leaders laud U.S. decision,” South China MorningPost, Business section 1994. May 28, 1994.
Topalova, Petia, “Trade liberalization, poverty, and inequality: Evidence from Indian dis-tricts,” in Ann Harrison, ed., Globalization and Poverty, Chicago: University of ChicagoPress, 2007, pp. 291–336.
, “Factor Immobility and Regional Impacts of Trade Liberalization: Evidence on Povertyfrom India,” American Economic Journal: Applied Economics, 2010, 2 (4), 1–41.
Wang, Zhi and Shang-Jin Wei, “What Accounts for the Rising Sophistication of China’sExports?,” in Robert C. Feenstra and Shang-Jin Wei, eds., China’s Growing Role in WorldTrade, Chicago: University of Chicago Press, 2010, pp. 63–104.
Zhang, Kevin H. and Shunfeng Song, “Promoting exports: The role of inward FDI inChina,” China Economic Review, 2000, 11, 385–396.
Zhang, Xiaobo, “Fiscal decentralization and political centralization in China: Implicationsfor growth and inequality,” Journal of Comparative Economics, 2006, 34 (4), 713–726.
50
A1 Appendix - for online publication
A1.1 Missing data
Data can be missing from the county-level panel for two reasons: counties cannot be matched
between the census and the provincial yearbooks, and counties are matched to the yearbooks
but specific indicators are not available. First, some counties that are observed in the census
do not appear in provincial yearbooks at all; these are disproportionately counties that are part
of the urbanized areas of larger, prefecture-level cities, as some provinces omit data for these
areas. To provide more systematic evidence about the characteristics of these missing counties,
we estimate the following specification at the county level in which county covariates as observed
in the 1990 census are regressed on a dummy variable equal to one if the county is missing due
to a failure to match between the census and the provincial yearbooks, as well as province fixed
effects. Standard errors are clustered at the province level.
Y 1990ifp = βMissingifp + ωp + εifp (9)
The results are reported in Table A1, employing six covariates: total population, the number
of households including children, the percentage of households including elderly, the percentage
of individuals who received a post-primary education, the percentage of individuals with an agri-
cultural registration, and the percentage of individuals working in the non-primary (secondary
and tertiary) sectors. We can observe that counties missing from the sample are characterized
by larger populations, higher levels of education, and a greater concentration of labor outside
of agriculture, consistent with the hypothesis that these are more urbanized counties.
Second, for those counties observed in provincial yearbooks, different provinces in different
years opt to report different indicators at the county level in their yearbooks. As a result, the
number of observations varies significantly for different variables, as evident from the summary
statistics. Thus (for example) Guangdong consistently reports employment data by sector for all
counties for all years; Shanxi does not report employment data by sector for any county in any
year. It is never the case that in the same yearbook (corresponding to a single province-year),
a particular indicator is reported for some counties and not for others. The indicators that are
reported most infrequently include employment at the sector level and exports, while indica-
51
tors reported near-universally include gross domestic product, total employment, population,
primary and secondary value added, and agricultural inputs and production.
Again, to provide more systematic evidence, we estimate the following specification, in which
the dependent variable is the number of observations for a particular variable observed for a
particular county, and the independent variables include the same county-level covariates as
observed in the 1990 census.
Obsifp = ξX1990ifp + ωp + εifp (10)
For counties that are missing, the number of observations is set at zero. The results are reported
in Table A2, and we again observe that the number of observations is in general lower for more
populous counties, and higher for those that are more agricultural and have a lower proportion
of employment outside the primary sector. This is consistent with more urban and industrialized
counties being generally underrepresented in the sample.
In the bottom of the table, we also report the average number of observations per county for
each variable, conditional on reporting any data. The main years represented in the sample are
the eighteen years from 1996 to 2013 inclusive, though a very small number of counties report
data for 1994 and 1995. The average number of observations per county is considerably lower
for all outcomes other than GDP and secondary value added, for which the average is 16-17
years per county. However, conditional on reporting any data, counties generally report at least
eight years of data for the key variables of interest, excluding only exports.
Finally, we turn to the question of whether selection into the sample is a source of bias
in the primary results. In Panel A of Table A3, we restrict the sample to county-years that
report export data. In Panel B, we include for each variable only the subset of counties that
reports at least eight observations for that variable, to avoid bias due to the entry and exit of
counties from the sample. In Panel C, we characterize each county and each variable according
to the number of observations, construct quartiles for the number of observations included, and
add observation quartile-year fixed effects. In all three cases, the results are robust. The only
exception is the coefficient for exports in Panel B, which is positive but not significant; however,
the sample has contracted dramatically in this specification. The consistency across a range of
specifications suggests that selection into the sample is not a significant source of bias.
52
Table A1: Counties missing from the provincial yearbooks
Pop. Hh incl. Hh incl. Post primary Prop. agri. Prop. non-prim.children elderly educ. empl.
(1) (2) (3) (4) (5) (6)
Missing counties 226877.700 -.0006 .010 .119 -.217 .229(63938.950)∗∗∗ (.007) (.005)∗ (.016)∗∗∗ (.052)∗∗∗ (.057)∗∗∗
Mean dep. var. 516645.34 .25 .21 .34 .83 .28Obs. 2104 2101 2101 2102 2100 2102
Notes: The dependent variables are variables reported in the 1990 census: county population, the percentage ofhouseholds including children age 0-7, the percentage of households including elderly members, the proportion of adultswith post-primary education, the proportion of the population designated as agricultural, and the proportion of thepopulation working in the secondary and tertiary sectors. The independent variable is equal to one if the county ismissing from the provincial yearbooks; all specifications include province fixed effects and standard errors clustered at theprovince level. Asterisks indicate significance at the ten, five and one percent level.
Table A2: Number of observations and initial county characteristics
Number of observationsGDP Primary Secondary Tertiary Export Emp. Secondary va(1) (2) (3) (4) (5) (6) (7)
Pop. -.002 -.0006 -.0006 -.0006 -.00004 -.0005 -.002(.0003)∗∗∗ (.0002)∗∗∗ (.0002)∗∗∗ (.0002)∗∗∗ (.0001) (.0001)∗∗∗ (.0003)∗∗∗
Prop. children 7.346 -1.216 .324 .315 -7.478 3.851 6.329(5.492) (3.084) (3.094) (3.050) (2.791)∗∗∗ (1.897)∗∗ (5.798)
Prop. elderly .934 .214 -.254 -.917 5.394 -.859 6.746(8.322) (4.673) (4.688) (4.622) (4.230) (2.874) (8.785)
Post.-primary 3.743 .829 1.996 1.992 7.702 -1.352 -.342(2.077)∗ (1.166) (1.170)∗ (1.153)∗ (1.056)∗∗∗ (.717)∗ (2.193)
Prop. agri. 5.533 3.226 .961 .783 7.247 1.000 7.125(2.159)∗∗ (1.213)∗∗∗ (1.216) (1.199) (1.098)∗∗∗ (.746) (2.280)∗∗∗
Prop. non-prim. -14.775 -4.395 -6.778 -7.118 -.941 -.189 -13.759(1.911)∗∗∗ (1.073)∗∗∗ (1.077)∗∗∗ (1.062)∗∗∗ (.971) (.660) (2.018)∗∗∗
Mean (obs. > 0) 16.52 8.31 8.36 8.33 5.07 8.3 17.12Obs. 2100 2100 2100 2100 2100 2100 2100
Notes: The dependent variable is the number of observations observed at the county level for the specified variable. Theindependent variables are a series of county characteristics observed at baseline: the fraction of the population engaged inprimary employment, the total population, and the fraction of the population with post-primary education (all observedin the 1990 census); GDP in the first year in which the county is observed in a provincial yearbook; and a dummy for anurban county. All specifications include prefecture fixed effects. Asterisks indicate significance at the ten, five and onepercent level.
53
Table A3: Main specifications controlling for selection into the sample
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: Sample reporting export data
Post x NTR gap 248.261 -8.361 268.547 128.211 431.175 1097.220(131.466)∗ (15.890) (91.398)∗∗∗ (58.774)∗∗ (169.113)∗∗ (293.562)∗∗∗
Mean dep. var. 751.069 1032.968 2582.935 1767.613 5863.137 9135.708Obs. 5105 2741 2761 2780 4959 4862
Panel B: Sample of counties including at least eight observations
Post x NTR gap 83.973 15.569 2479.892 1240.035 1244.824 1298.070(118.063) (32.449) (705.281)∗∗∗ (1040.152) (650.375)∗ (389.925)∗∗∗
Obs. 3060 10800 10883 10754 27983 26072
Panel C: Including observation number quartile - year fixed effects
Post x NTR gap 238.277 39.231 4018.797 1581.600 2811.291 936.339(130.700)∗ (34.705) (844.659)∗∗∗ (980.363) (728.368)∗∗∗ (390.499)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate). Province-year and county fixed effectsand prefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standarderrors are estimated employing clustering at the county level. The dependent variables include exports at the countylevel; primary, secondary, tertiary and total GDP; and per capita GDP.
In Panel A, the sample is restricted to county-year observations reporting export data. In Panel B, the sample for eachvariable is restricted to the subset of counties that report at least eight observations for this variable. In Panel C, wecharacterize each county by the number of observations reported for each variable, and construct quartiles for the numberof observations reported; the specification then includes quartile-year fixed effects. Asterisks indicate significance at theten, five and one percent level.
54
A1.2 Calibrations for aggregate effects
In order to conduct a simple back-of-the-envelope calculation estimating the contribution of
the export shock captured by the NTR gap to total growth over this period, we must first
make some assumptions about the magnitude of the decrease in uncertainty experienced by the
average county between the pre and the post period. In theory, uncertainty captured by the
NTR gap is reduced to zero in the post-WTO period. In practice, some residual uncertainty
may have been perceived. In addition, given that the minimum level of the NTR gap observed
in the pre-WTO period is .064, calculating effects for a reduction in the NTR gap to zero is an
out-of-sample calculation that may be hard to defend. Accordingly, we assume that the median
county experiences a reduction in the NTR gap to the minimum level of the gap observed ex
ante, or a decrease of one standard deviation in the NTR gap.
We can then calculate the implied increase in GDP (32%), secondary GDP (72%), and
secondary employment (11%) predicted for the median county in the post-WTO period, and
compare these implied increases to the median overall growth in each measure observed in the
sampled counties in this period. We calculate the median growth rates across all counties that
report the indicators of interest; the median county shows growth in total GDP of 227%, in
secondary GDP of 201%, and in secondary employment of 78%. The ratio of the first measure
to the second is the proportion of total growth generated by the reduction in tariff uncertainty.
For employment, we additionally convert this estimate of growth into an estimate of the
absolute increase in employment, using median secondary employment observed at the county
level in the 2000 census and focusing on the 1805 counties observed in the sample. Since
secondary employment is not universally observed in 2002, we impute median employment in
2002 (the first post-WTO year) using the 2000 census data and the average annual growth
rate in secondary employment observed in the sample. The number of new positions created
by the reduction in tariff uncertainty post-WTO in the sampled counties is then calculated
using 2002 estimated employment and the estimated growth rate of 11%.44 Given that the
sampled counties account for 74.7% of total secondary employment in the 2000 census, with
the remaining quarter attributable to counties not observed in the sample, we then adjust this
estimate upward accordingly.
44More specifically, the calculation is 48397∗.11∗1805, where 48397 is estimated average secondary employmentper county in 2002, .11 is the estimated growth generated by the NTR shock, and 1805 is the number of counties.
55
It should be noted that these estimates reflect a number of underlying assumptions and
should be considered merely illustrative. However, they do suggest that the effects of export
expansion driven by the reduction in uncertainty are non-trivial in magnitude.
A1.3 Night lights data
To generate county-level averages of night lights data, high-resolution data on light density mea-
sured by satellites at night was downloaded from http://ngdc.noaa.gov/eog/dmsp/downloadV4composites.
html. ArcMAP was used to process the files. They were converted from raster to GRID files,
and averaged across the two sets of data available for each year.45
Next, gas flares were removed by using the China gas flares shapefile, downloaded from
https://www.ngdc.noaa.gov/eog/data/web_data/gasflares_v2/country_vectors_20090618/
Flares_China_1.tgz. The erase function in ArcMap was used to remove the gas flares from
the dataset. The average night light intensity observed within each county boundary was then
calculated using the zonal statistics function and the shape file corresponding to the 2010 county
boundaries. The night lights variable ranges from 0 to 63 across counties for each year, based
on the highest and lowest brightness captured on the satellite images.
45This excludes only 2007, given that only one dataset is available.
56
A1.4 Appendix Figures and Tables
Figure A1: Variation in Tariff Policy Across Counties and Over Time
(a) China’s Import Tariffs Over Time
(b) Major Trading Partners’ Tariffs Over Time
Notes: The first subfigure shows the average domestic import tariff and the second subfigure shows the mean tariffsimposed on Chinese exports by major trading partners from 1996 to 2013. For each variable, we calculate the averagecounty-level weighted average tariff, using tariffs by industry and employing as weights the county-level employment shareof each industry as reported in the 1990 census. We then report the mean weighted tariff over all counties in each year.Tariff data is obtained from the WITS-TRAINS database.
57
Figure A2: NTR Gap by County
Quantile 1Quantile 2Quantile 3Quantile 4Non-sample areas
Notes: This figure shows the NTR gap at the county level, utilizing the residuals from the gap regressed on provincialfixed effects. Areas not shaded are out of sample. This includes the autonomous regions (Inner Mongolia, Guangxi,Ningxia, Tibet, and Xinjiang) and counties that cannot be matched between the county-level census data and theprovincial yearbooks.
58
Figure A3: China’s Agricultural Imports from the U.S.
Notes: This figure shows the evolution of Chinese imports of agricultural products from the U.S. during the period ofinterest in the primary analysis.
59
Table A4: Concordance between Chinese Census Industry Categories, ISIC andSIC
Chinese Census Industry: ISIC Revision 3: SIC:Codes Labels 2-Digit 3-Digit 2-Digit 3-Digit
90136 Farming 11 190137 Forestry 2 890138 Animal Husbandry 12 290139 Fishery 5 990140, Agricultural Services 11 79014190142 Coal Mining and Dressing 10 1290143 Extraction of petroleum and Natural Gas 11 1390144 Mining and Dressing of Ferrous Metals 12 10190145 Mining and Dressing of Nonferrous Metals 13 102, 103, 104, 105,
106, 107, 108, 10990146 Mining and Dressing of Nonmetal Minerals 141 14190147 Mining and Dressing of Other Minerals 142 142, 143, 144, 145,90148 146, 147, 148, 14990149 Logging and Transport of Wood and Bamboo 2 24190151 Food Processing 151, 152, 201, 202, 203, 204,
153, 154 205, 206, 207, 20990152 Beverages 155 20890153 Tobacco 16 2190155 Textiles 17 2290156 Garments and Other Fiber Products 18 2390157 Leather, Furs, Down and Related Products 19 3190158 Timber Processing, Bamboo, Cane, 20 24
Palm Fiber and Straw Products90159 Furniture Manufacturing 361 2590160 Paper-making and Paper Products 21 2690161 Printing and Record Medium Reproduction 22 2790165, Petroleum Processing and Coking 23 299016690167 Raw Chemical Materials and Chemical Products 241, 242 281, 283, 284, 285,
286, 287, 288, 28990168 Medical and Pharmaceutical Products 33 38490169 Chemical Fiber 243 28290170 Rubber Products 251 301, 302, 303, 304,
305, 30690171 Plastic Products 252 30890172 Nonmetal Mineral Products 26 3290173 Smelting and Pressing of Ferrous Metals 271 331, 33290174 Smelting and Pressing of Nonferrous Metals 272 333, 334, 335, 336,
337, 338, 33990175 Metal Products 28 341, 342, 343, 344,
345, 346, 347, 34990176 Ordinary Machinery 291, 293 351, 352, 353, 35490177 Transport Equipment 34, 35 3790178 Electric Equipment and Machinery 31 361, 362, 363, 364,
36590179 Electronic and Telecommunications Equipment 32 366, 367, 368, 36990180 Instruments, Meters, Cultural, 30 38
and Office Machinery90181 Other Manufacturing 369 39
Notes: This table reports the industry categories and their labels in the 1990 Chinese Census that can be matched toISIC Revision 3 codes and SIC codes. Three-digit codes represent more disaggregated industry categories compared totwo-digit codes. All industry categories reported in the Chinese Census are matched to two- or three-digit codes in ISICor SIC codes. The category of cultural, educational, and sporting goods (90162, 90163) does not match to the ISIC orSIC codes, and is therefore excluded.
60
Table A5: NTR gap by industry
Subsectors NTR gap
Coal Mining and Dressing .000Mining and Dressing of Ferrous Metals .000Fishery .012Extraction of petroleum and Natural Gas .059Mining and Dressing of Nonferrous Metals .061Animal Husbandry .076Petroleum Processing and Coking .088Farming .096Agricultural Services .096Forestry .123Logging and Transport of Wood and Bamboo .123Mining and Dressing of Other Minerals .128Food Processing .134Mining and Dressing of Nonmetal Minerals .175Smelting and Pressing of Ferrous Metals .199Beverages .201Timber Processing, Bamboo, Cane, Palm Fiber and Straw Products .206Rubber Products .217Transport Equipment .222Smelting and Pressing of Nonferrous Metals .231Printing and Record Medium Reproduction .242Raw Chemical Materials and Chemical Products .269Leather, Furs, Down and Related Products .283Papermaking and Paper Products .284Cultural, Educational and Sports Goods .305Nonmetal Mineral Products .309Tobacco .317Instruments, Meters, Cultural and Office Machinery .321Electric Equipment and Machinery .334Electronic and Telecommunications Equipment .338Ordinary Machinery .363Metal Products .3835Chemical Fiber .383Plastic Products .420Furniture Manufacturing .424Medical and Pharmaceutical Products .425Other Manufacturing .426Garments and Other Fiber Products .457Textiles .523
Notes: This table reports the NTR gap by industry for each tradable subsector reported in the 1990 Chinese countycensus.
61
Table A6: Exports and GDP, Full Results
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Post x NTR gap 248.261 42.915 4613.042 1875.907 2655.506 955.651(131.466)∗ (32.992) (1254.208)∗∗∗ (870.109)∗∗ (744.508)∗∗∗ (400.357)∗∗
Post x Contract -19.836 -34.934 -3892.620 -1511.175 -1978.065 369.808(110.808) (31.414) (1033.102)∗∗∗ (796.223)∗ (568.893)∗∗∗ (429.311)
MFA -175.185 -18.735 1407.584 972.672 1089.513 866.114(239.296) (12.358) (638.449)∗∗ (412.872)∗∗ (624.229)∗ (277.518)∗∗∗
Tariff -135.534 64.794 8131.321 10534.280 6086.603 2769.207(706.651) (67.326) (2823.478)∗∗∗ (3344.719)∗∗∗ (1803.809)∗∗∗ (852.823)∗∗∗
Post x Tariff -849.507 -25.902 -3926.651 6928.947 -2661.801 -2633.559(588.077) (140.953) (1623.368)∗∗ (3834.144)∗ (1974.211) (1453.923)∗
License 57.074 3.605 -289.058 414.795 -793.522 -1609.757(82.422) (31.002) (666.677) (465.293) (446.943)∗ (396.940)∗∗∗
NTR rate 412.644 -38.980 597.889 173.387 322.218 510.778(110.401)∗∗∗ (34.966) (481.107) (387.893) (291.409) (178.350)∗∗∗
Mean dep. var. 752.808 1010.083 6371.597 4642.585 8332.215 9666.646Counties 1005 1465 1458 1455 1655 1595Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level.
The dependent variables include exports at the county level; primary, secondary, tertiary and total GDP; and per capitaGDP. Exports and GDP are reported in millions of yuan deflated to 2000 constant prices; per capita GDP is reported inyuan. Asterisks indicate significance at the ten, five and one percent level.
62
Table A7: Employment, Full Results
Primary Secondary Tertiary Agri. Total emp. Total pop.(1) (2) (3) (4) (5) (6)
Post x NTR gap -5.006 7.678 -1.327 -7.860 -.557 35.040(6.941) (4.784) (3.122) (4.220)∗ (2.713) (12.241)∗∗∗
Post x Contract 3.585 -1.363 3.447 -.558 -5.426 -23.229(6.259) (4.698) (3.029) (2.677) (3.249)∗ (9.221)∗∗
MFA -3.102 1.150 -1.964 8.859 2.870 -5.077(4.938) (2.070) (1.769) (3.980)∗∗ (1.687)∗ (4.161)
Tariff 3.036 3.911 -3.248 -4.067 1.299 -17.887(4.961) (3.436) (2.577) (2.506) (5.296) (21.065)
Post x tariff 11.917 -4.693 -.294 18.459 -5.241 -84.198(10.032) (6.921) (6.449) (6.655)∗∗∗ (7.521) (40.378)∗∗
License 1.336 2.035 1.472 -3.255 -7.936 -7.601(6.477) (4.666) (5.167) (3.083) (3.135)∗∗ (4.228)∗
NTR rate .663 -.792 .123 -2.795 .836 -15.835(1.662) (1.375) (1.051) (1.137)∗∗ (1.987) (14.339)
Mean dep. var. 173.968 66.366 72.675 217.098 250.002 509.042Counties 353 398 399 1305 1419 1602Obs. 3127 3376 3501 20634 19210 27802
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level.
The dependent variables include employment in the primary, secondary and tertiary sectors, total employment, andpopulation; all variables are reported in thousands of persons. Asterisks indicate significance at the ten, five and onepercent level.
63
Table A8: Agricultural Investment and Value added, Full Results
Sown area Agri. machine Grain Cash Primary va Secondary va(1) (2) (3) (4) (5) (6)
Post x NTR gap .152 -3.068 -8.078 -1.400 -2.615 51.358(.964) (1.585)∗ (4.772)∗ (1.007) (1.463)∗ (19.680)∗∗∗
Post x Contract -2.192 1.628 10.215 2.606 3.836 -7.542(2.020) (1.315) (4.978)∗∗ (1.366)∗ (1.958)∗ (18.672)
MFA -.192 -.344 -6.636 -.145 -.984 28.220(.555) (.666) (1.264)∗∗∗ (.489) (.589)∗ (19.265)
Tariff 1.019 -9.170 -51.022 -4.213 -9.922 104.800(3.871) (2.310)∗∗∗ (9.900)∗∗∗ (1.457)∗∗∗ (2.332)∗∗∗ (21.434)∗∗∗
Post x Tariff -3.703 7.539 -62.175 12.410 2.275 -64.288(7.274) (5.530) (20.595)∗∗∗ (3.256)∗∗∗ (4.899) (39.796)
License 2.465 -.963 -1.143 -1.336 3.652 -75.757(1.401)∗ (1.357) (4.782) (1.683) (2.008)∗ (15.349)∗∗∗
NTR rate .516 .361 -5.612 -1.506 -.432 19.224(.867) (.917) (3.330)∗ (.607)∗∗ (.851) (7.648)∗∗
Mean dep. var. 61.54 33.838 230.222 42.341 86.375 223.996Counties 966 1621 1600 1559 1574 1560Obs. 8019 27277 27184 25779 27054 27126
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. Standard errorsare estimated employing clustering at the county level.
The dependent variables include sown area in thousands of hectares, agricultural machinery employed in 10,000 kilowatts,grain and cash crops in thousands of tons, and primary and secondary value added in millions of yuan deflated to 2000constant prices. Asterisks indicate significance at the ten, five and one percent level.
64
Table A9: Alternate Trimming
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: Baseline specification without trimming
Post x NTR gap 566.767 39.720 4402.248 1417.568 4340.623 736.496(374.532) (33.277) (1249.551)∗∗∗ (1010.877) (1458.594)∗∗∗ (514.382)
Obs. 5360 13533 13544 13444 29508 27340
Panel B: Baseline specification with trimming at the 1/99 percentiles
Post x NTR gap 749.983 39.711 4397.729 1416.080 4337.329 736.496(293.664)∗∗ (33.263) (1249.808)∗∗∗ (1010.426) (1458.913)∗∗∗ (514.382)
Obs. 5152 13231 13242 13094 28756 26654
Panel C: NTR Gap Winsorized at 1/99 percentile
Post x NTR gap 559.012 42.859 3767.659 1126.670 2310.271 1240.756(222.371)∗∗ (29.625) (1276.323)∗∗∗ (787.331) (820.626)∗∗∗ (423.256)∗∗∗
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. The dependentvariables are exports at the county level; primary, secondary, tertiary and total GDP; and per capita GDP. Standarderrors are estimated employing clustering at the county level.
In Panel A, the dependent variables are not trimmed; in Panel B, observations above (below) the 99th (1st) percentile ineach year for urban and non-urban counties respectively are trimmed. (The trimming procedure does not eliminateobservations with a value of zero if the first percentile of the variable is zero; for this reason, the number of exportobservations dropped is relatively low, given that the first percentile of exports is zero in a number of cells.) In Panel C,the original sample is employed for the dependent variables, and the NTR gap is winsorized at the 1st and 99thpercentiles. Asterisks indicate significance at the ten, five and one percent level.
65
Table A10: Alternate NTR Gaps
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: NTR gaps estimated excluding the agricultural sector
Post x NTR gap 1521.185 81.810 19278.950 10249.430 12455.450 4149.234(467.125)∗∗∗ (89.186) (5016.792)∗∗∗ (3505.490)∗∗∗ (2722.344)∗∗∗ (1149.468)∗∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel B: NTR gaps estimated excluding high-gap subsectors
Post x NTR gap 403.751 20.852 4799.112 2256.003 2620.819 851.515(165.998)∗∗ (21.219) (1474.057)∗∗∗ (983.504)∗∗ (780.454)∗∗∗ (406.433)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. The specification also includes a number of control variables: an interaction of the post-reform indicatorvariable and a time-invariant dummy capturing whether the county is characterized by high contract intensity industries,the industry-weighted MFA quota fill rate for county-produced goods, the industry-weighted national tariff rate forimports of county-produced goods and the tariff rate interacted with the post dummy, the industry-weighted percentageof firms licensed to export, and the industry-weighted time-varying NTR rate. Province-year and county fixed effects andprefecture-specific trends are included, and the time-varying trends are interacted with an urban dummy. The dependentvariables include exports at the county level; primary, secondary, tertiary and total GDP; and per capita GDP. Standarderrors are estimated employing clustering at the county level.
In Panel A, the NTR gap is re-estimated excluding the agricultural sector. In Panel B, the NTR gap is re-estimatedexcluding the five subsectors (textiles, garments, other manufacturing, medical and pharmaceutical products, andfurniture manufacturing) characterized by the highest NTR gaps. Asterisks indicate significance at the ten, five and onepercent level.
66
Table A11: Alternate Specifications
Exports Primary Secondary Tertiary GDP Per capita(1) (2) (3) (4) (5) (6)
Panel A: Baseline specification without controls
Post x NTR gap 233.762 106.314 4390.154 2956.292 4876.070 2059.562(110.911)∗∗ (30.109)∗∗∗ (634.128)∗∗∗ (855.871)∗∗∗ (1054.134)∗∗∗ (400.849)∗∗∗
Obs. 5133 13551 13570 13441 28942 26595
Panel B: Population-weighted standard errors
Post x NTR gap 448.531 18.804 5107.279 2363.045 3177.225 526.810(233.164)∗ (45.259) (1608.439)∗∗∗ (1122.092)∗∗ (1030.619)∗∗∗ (413.230)
Obs. 5102 13028 13047 12916 28340 26301
Panel C: Baseline GDP quartile - year fixed effects
Post x NTR gap 244.223 43.991 4632.707 1878.242 2614.256 919.508(130.629)∗ (33.322) (1261.550)∗∗∗ (874.446)∗∗ (753.774)∗∗∗ (403.419)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel D: Baseline education quartile - year fixed effects
Post x NTR gap 284.565 46.911 4602.560 1818.345 2711.616 1008.400(131.891)∗∗ (32.906) (1236.040)∗∗∗ (893.673)∗∗ (745.461)∗∗∗ (404.710)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel E: Baseline concentration - year fixed effects
Post x NTR gap 350.467 48.444 4901.302 2026.189 3046.716 1096.813(148.868)∗∗ (34.204) (1328.597)∗∗∗ (915.199)∗∗ (786.442)∗∗∗ (430.659)∗∗
Obs. 5105 13034 13053 12922 28355 26316
Panel F: SOE employment-quartile fixed effects
Post x NTR gap 318.807 45.442 5004.815 2031.854 2834.238 1152.447(132.080)∗∗ (33.102) (1332.208)∗∗∗ (913.927)∗∗ (725.690)∗∗∗ (407.144)∗∗∗
Obs. 5105 13034 13053 12922 28355 26316
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates, standardized to have mean zero and standarddeviation one. In all panels except Panel A, the control variables and fixed effects included are identical to those reportedin the notes to Table 2. The dependent variables include exports at the county level; primary, secondary, tertiary andtotal GDP; and per capita GDP.
In Panel A, the specification is estimated including only county and province year fixed effects and prefecture-specifictrends. In Panel B, the observations are weighted with respect to the 1990 county population. In Panel C, a full set ofinteractions between year fixed effects and dummy variables for each quartile of initial GDP are added. In Panel D, a fullset of interactions between year fixed effects and dummy variables for each quartiles of initial post-primary education areadded. In Panel E, a full set of interactions between year fixed effects and dummy variables for each quartile of the initialHerfindahl index are added. In Panel F, a full set of interactions between year fixed effects and dummies for each quartileof estimated baseline fraction of SOE employment are added. Asterisks indicate significance at the ten, five and onepercent level.
67
Table A12: Foreign Direct Investment
County-level data Provincial-level dataContractual Foreign capital Total foreign Foreign loans Direct FDI
FDI used capital(1) (2) (3) (4) (5)
Post x NTR gap 31.552 109.961 1027.118 261.453 850.359(30.316) (47.449)∗∗ (505.318)∗∗ (102.664)∗∗ (661.606)
Mean dep. var. 133.463 120.757 2354.97 336.663 2108.981Obs. 5436 5540 279 213 376
Notes: In Columns (1) and (2), the dependent variables are contractual FDI and foreign capital used calculated inmillions of yuan reported at the county level. The primary independent variables are the interaction of a dummy variableequal to one for the post–2001 period and the county-level gap between NTR tariffs and the non-NTR rates, standardizedto have mean zero and standard deviation one. The specification also includes a number of control variables: aninteraction of the post-reform indicator variable and a time-invariant dummy capturing whether the county ischaracterized by high contract intensity industries, the industry-weighted MFA quota fill rate for county-produced goods,the industry-weighted national tariff rate for imports of county-produced goods and the tariff rate interacted with thepost dummy, the industry-weighted percentage of firms licensed to export, and the industry-weighted time-varying NTRrate). Province-year and county fixed effects and prefecture-specific trends are included, and the time-varying trends areinteracted with an urban dummy. Standard errors are estimated employing clustering at the county level.
In Columns (3) through (5), a parallel specification is estimated at the province-year level. The NTR gap and all othercontrol variables are calculated as the mean variable across counties in the specified province and year. Province and yearfixed effects are included, and standard errors are estimated employing clustering at the province level. Asterisks indicatesignificance at the ten, five and one percent level.
68
Table A13: Firm Outcomes: Specifications Using Weighted Means
Panel A: Factor utilization
Emp. Capital Foreign capital Cap. Wages Wages perdummy intensity worker
(1) (2) (3) (4) (5) (6)
Post x NTR gap 44.543 1356.890 .010 -2.029 950.066 .264(19.514)∗∗ (432.914)∗∗∗ (.006)∗ (1.921) (321.223)∗∗∗ (.148)∗
Mean dep. var. 116.35 3655.667 .077 61.848 1354.195 9.416Obs. 2423 2020 2446 1997 2446 2423
Panel B: Other firm outcomes
Exports Sales Value added Profits Value-added per worker
Post x NTR gap 2747.186 17025.090 3979.107 1662.164 .561(1062.459)∗∗∗ (4553.016)∗∗∗ (1027.445)∗∗∗ (379.734)∗∗∗ (1.873)
Mean dep. var. 3354.639 24658.011 5212.273 2143.398 51.59Obs. 2446 2446 2231 2446 2207
Notes: The primary independent variable is the interaction of a dummy variable equal to one for the post–2001 periodand the county-level gap between NTR tariffs and the non-NTR rates at the prefecture level, standardized to have meanzero and standard deviation one. The specification includes the same control variables described in the notes to Table 2,all calculated as the average at the prefecture-year level, as well as province-year and prefecture fixed effects. Allstandard errors are estimated employing clustering at the prefecture level.
The dependent variables in Panel A include total employment in sampled firms, the total wage bill in sampled firms,mean wage per worker, total capital stock in sampled firms, and mean capital intensity. The dependent variables in PanelB include total exports, sales, value added and profits in sampled firms, as well as mean value added per worker.Asterisks indicate significance at the ten, five and one percent level.
69